Google ($GOOG) dropped another monster quarterly report earlier this week, reporting $61.9 billion in revenue for Q2 2021 and $18.5 billion in net income. Net income was particularly impressive, in that it was more than double the $6.96 billion reported one year ago.

The vast majority of that amount came from Google’s services division, which includes online advertising, sales of apps and hardware, and subscription fees for various products. But tucked away in the footnotes, we noticed one additional gem: Google has extended the estimated useful life of its servers and other equipment, which added an extra $561 million to the bottom line in Q2.

Hey, every little bit helps, right?

The full disclosure is listed in Google’s Summary of Significant Accounting Policies, which you can find using Calcbench’s Interactive Disclosure viewer. There, under the headline “Change in Accounting Estimate,” the company says this:

In January 2021, we completed an assessment of the useful lives of our servers and network equipment and adjusted the estimated useful life of our servers from three years to four years and the estimated useful life of certain network equipment from three years to five years … The effect of this change in estimate was a reduction in depreciation expense of $721 million and $1.6 billion and an increase in net income of $561 million and $1.2 billion, or $0.84 and $1.81 per basic and $0.83 and $1.78 per diluted share, for the three and six months ended June 30, 2021, respectively.

Put more simply: Google decided its equipment can last longer than originally expected. That pushed down the company’s depreciation expense, and the savings fell directly into the net income line.

Of course, $561 million is only 3 percent of Google’s total net income for the quarter, which barely qualifies as material. We were more interested in the nature of the accounting change rather than its size — because fiddling with the estimated useful life of assets is something any company can do.

Our columnist Jason Voss, for example, recently wrote about the Property, Plant & Equipment line item and how an analyst might estimate the average life of fixed assets. As Voss noted in his column, “Many firms play games with depreciation & amortization to improve their short-term results, and such games depend on PP&E.”

To be clear, we are not suggesting that Google is playing with estimated asset lifespan here. Other firms, however, have been called out for such abuses. For example, in 2019 the Securities and Exchange Commission hammered Hertz on accounting fraud. One of the company’s schemes was to extend the estimated life of its rental fleet from 20 months (the industry norm) to 30 months, which kept depreciation costs low and therefore goosed net income up.

So Google’s morsel of disclosure is just another reminder that analysts should always be on the lookout for details that might tell more than you’d expect — and as always, Calcbench has the data!

Monday, July 26, 2021

With all the talk about SPACs (special purpose acquisition companies) and “deSPACing” (when a SPAC acquires an operating company and the two fuse into one publicly traded firm), some people may feel like everyone is doing it these days.

So, continuing with our ongoing series about SPACs, we decided to check. We wanted to see how prevalent SPACs and deSPACing really are.

Figure 1, below, tells the tale. We charted the number of SPACs each year (orange line) for the last decade, compared to the number of firms that deSPACed (blue line) in the same year.

How did we get those numbers? SPACs are defined as firms that use SIC 6770, the code for blank check companies; and a deSPACing is detected by a change in that reported SIC from 6770 to something else. These cases are easy to identify as Calcbench has a metric called “SIC Code at Time of Filing.”

As one can see from the chart, the number of SPACs is not at an all-time high, even though it feels as such these days. In fact, however, the number of SPACs today is just a bit lower than SPACs in 2011.

What is striking, however, is the number of firms that have deSPACed. The number of deSPACss was much higher in 2020 (85 cases) compared to previous years. We’re on pace to be at least that high in 2021, with 42 deSPACs already reported through the first half of this year.

We then examined total assets reported by these firms. Figure 2, below, shows total assets for SPACs and the resulting firms after a deSPAC transaction.

So if we take Figures 1 and 2 together, we can see that while there were more SPACs in 2011 and 2012 than there are today, those SPACs didn’t have much in total assets. Today’s SPACs have total assets north of $190 billion.

Although there was a high number of SPACs in 2011 and 2012, their total assets did not add up to much. With the more recent increase in the number of SPACs, we see a substantial increase in total assets reaching a total of $190 billion. Total assets for firms after de-SPACing (shown using the left-side axis in Figure 2) also increased dramatically, to $34 billion.

Of course, $34 billion in deSPACed assets is still a lot less than the $190 billion in assets SPAC firms have in their coffers. So presumably we’re going to see a lot more of that money flow into merger deals and deSPAC transactions, as the SPACs come under pressure to put all that money to work.

Thursday, July 22, 2021

We have one more post about SPACs today to explore the final phase of existence for these firms: what happens to a special purpose acquisition company after it acquires an operating company. Or, more simply — where do SPACs go after they die?

A fair number of them go into the software and tech sectors, apparently.

The crack Calcbench research team examined 135 SPACs to see what operating businesses they acquired, and what the balance sheets of those resulting firms looked like. Of those 135 firms, 22 “de-SPACed” into prepackaged software businesses and another 16 went into computer processing and data preparation.

Following a SPAC firm into the operating company beyond isn’t easy, because once the SPAC acquires the target, it ceases to exist. No longer do you have a ticker symbol or other standard identifier that would let an analyst follow the SPAC from its prior form into the new one.

But as a matter of standard procedure, Calcbench does assign a unique identifier to every registrant in our database, SPACs included. So we can follow along and see which operating companies today were SPACs in the past. By examining the resulting firm’s new SIC code, we can see which industries were most popular with SPACs as the SPACs acquired their way into operating company status.

Figure 1, below, is the breakdown for the 135 SPACs we studied. As you can see, the picture quickly gets messy. After prepackaged software, business services, and computer processing, most other industries have only a handful of SPACs entering their lines of work.

Where the Money Goes

We were also curious to see how de-SPAC transactions work in specific instances. So we looked at Diamond Eagle Acquisition Corp., which acquired both online sports betting company DraftKings ($DKNG) and sports betting technology firm SBTech at the same time in April 2020.

Prior to the acquisitions, Diamond Eagle had about $400 million in its coffers to make an acquisition. Executives there struck a deal to acquire DraftKings and SBTech at the end of 2019, and part of the deal included several institutional investors pumping another $304 million into the resulting public company in exchange for stock. That’s known as a PIPE deal (private investment in public equity) and they’re a common part of de-SPAC transactions.

The accounting for de-SPAC transactions, however, is tricky. Essentially, accounting rules treat DraftKings as the acquiring firm, which scooped up both Diamond Eagle and SBTech. So there is no purchase price allocation one can examine to see what the SPAC paid for DraftKings.

Rather, Diamond Eagle and the PIPE investors pumped a total of $704 million into DraftKings and SBTech, and the resulting public company had $1.82 billion in cash on the balance sheet at the end of 2020. (Up from only $76.5 million in 2019.)

On the other hand, DraftKings also acquired that other business, SBTech — and we do have a purchase price allocation for that one. If you look at the Business Combinations disclosure that DraftKings included in its 10-K from May 3, you find this:

In other words, DraftKings (or Diamond Eagle and its PIPE investors, depending on your point of view) paid $977 million to acquire SBTech, and $538 million of that sum (55 percent) went to goodwill.

You can conduct similar research yourself using our Interactive Disclosures page to research business combinations, with some assistance from the Company in Detail page to analyze the balance sheet.

And if you want more information about tracing SPACs from their pre-acquisition life to post-acquisition existence as a new operating company, using our Calcbench identifier, just drop us a line at and we’re happy to help you out.

Monday, July 19, 2021

Editor’s note: Calcbench recently published a column from Jason Apollo Voss on analyzing EPS and price-to-earnings metrics. Allison Kays, assistant professor of accounting at Emory University, wrote the following rebuttal. We welcome her feedback and post her analysis in its entirety.

By Allison Kays, PhD CPA

In his July 6 column titled, “Is it Really Earnings, or Just Financial Engineering?” Jason Apollo Voss makes the important point that readers of financial statements always need to think critically when interpreting ratios. After starting with ratio analysis, one should dig deeper into a firm’s financial statements to determine why a ratio increased or decreased over time or is higher or lower than a competing firm’s ratio.

Voss’s column  focuses on earnings per share (EPS) and its susceptibility to manipulation  by managers. I wholeheartedly agree and would even go so far as to say that on its own, EPS is a completely useless ratio. Which is especially interesting since EPS is the only ratio that is explicitly defined by U.S. GAAP and required to be reported within a firm’s financial statements.

The reason that EPS is useless on its own is because a firm’s number of common shares outstanding is arbitrary. Two firms of the same size (similar assets and revenue), profitability, and level of common equity funding can have completely different numbers of common shares outstanding. That will result in the two firms reporting completely different EPS ratios. This is because the number of shares outstanding is a measure of how many slices we cut the pie into (which is, of course, arbitrary) rather than a measure of how big the pie is.

To make EPS a more informative metric, it is often compared to the price of a share of a firm’s common stock in a price-to-earnings ratio:

The great thing about the P/E ratio is that both the numerator and the denominator are per share numbers; that allows the fraction to become independent of the number of shares outstanding. To further emphasize this fact, the ratio can also be calculated at the firm level (rather than per share):

In his column, Voss argues that the latter version of the ratio is superior to the former version. It’s important to realize, however, that the two ratios are identical. That can be proven with a simple proof:

Dividing by a fraction is the same as multiplying by its reciprocal:

This proof begs the question: why does Voss’s calculation differ from a traditional P/E ratio when he applies it to S&P 500 companies?

The answer lies in the details. As I mentioned before, EPS is the only ratio explicitly defined by U.S. GAAP and reported on a firm’s financial statements. U.S. GAAP defines EPS as:

My guess is that the difference in Voss’s ratios lies in using the number of shares outstanding at year-end rather than the weighted average number of shares outstanding in the traditional calculation. (A footnote: Very few firms have preferred dividends, but this adjustment could also make a difference in your calculation. It is important to subtract out preferred dividends as that portion of income is not earned by common shareholders.)

The weighted average number of shares outstanding calculates the average shares outstanding over the course of the year, weighted by the length of time the shares are outstanding.

One might argue that since price is a year-end number, we should use the number of shares outstanding at year-end. That’s perfectly fine as long as you understand the choice you are making and its implications.

The reason that U.S. GAAP requires the use of a weighted average is because earnings are earned over a period of time. To make the numerator and the denominator comparable, the weighted average number of shares outstanding is designed to capture the common equity funding available over the same period of time.

The logic here is that when a firm issues more shares, it receives more funding — funding which will likely be used to buy income-producing assets which increases the firm’s income earning ability. On the other hand, if a firm buys back stock, it uses assets to buy that stock back, which makes the firm smaller and reduces the firm’s ability to earn income. (Another footnote: Even if the company uses new funding to pay off debt (or uses debt to fund a share buyback), interest expense will go down (up) which will increase (decrease) net income.)

If the shares are issued or bought back halfway through the year, the company’s ability to earn income is only affected for half of the year. Thus, the denominator takes this effect into account by weighting the shares outstanding accordingly. (One more footnote: This method also makes it harder for a manager to make a last minute manipulation of EPS. If the manager buys back shares right at year end to inflate EPS, there will be very little weight on this reduction to shares outstanding as it only existed for 1/365 days of the year.)

Voss finds that the S&P 500’s P/E ratio is consistently higher when calculated with the number of shares outstanding at year-end rather than using the weighted average. This suggests that on average, firms have more shares outstanding at year-end than over the course of the year.

Lastly, in Part II of Voss’column, he argues that comparing growth in net income over time versus growth in EPS over time is a measure of the level of managerial manipulation of EPS.  It’s important, howevever, also to consider the relationship I describe above: if a company issues more shares, it receives more capital. If it receives more capital, it should use that capital to invest in incom- producing assets, which should produce more income.

Thus, you would expect a company with growing shares outstanding also to have growing income, because the firm is growing. If you look at growth in net income you are measuring growth without considering the amount of funding provided by equity holders.

Voss shows that Google’s net income is growing more quickly than its EPS. This comparison suggests that the company is not using its new funding as efficiently. Voss also shows that Apple’s net income is growing more slowly than its EPS. This comparison suggests that even though the company is buying back shares (spending money on a nonproductive investment) it is still able to grow its business. The key is to understand why the ratio changes, by looking at changes in both the numerator and denominator rather than making any assumptions.

In conclusion, there is not one right way to define any ratio. The internet is full of different definitions of the same ratio. That said, it’s also important to understand the different choices and their implications so that the ratios can be properly interpreted.

A reply from Jason Voss: I love this, and that the author took the time to poke holes in my thinking! As financial due-diligence professionals it is important that we constantly be on a quest to get better. I think that Dr. Kays’ post helps all of us improve our understanding of financial statements, and that is a very good thing. Well done.

Did you know that you can do a few simple things to use text analytics in Calcbench?  Of course you know about the search functions of our website which allow you to search subsections of documents like various disclosures, aka footnotes.  You can pinpoint search only disclosures that you want, so you could, for example, look at a handful of Fair Value disclosures for a subset of banks and ignore the rest of the document.  

Today, we are going to extend the concept of disclosure search by using our Excel Add-In to look at text from Earnings Press Releases.  The spreadsheet here allows you to pinpoint search for a term within an Earnings Press Release and count the number of instances of the item.  

Specifically, this case searches a press release for the Frequency of the word “Inflation”.  You can also compare the frequency of the term to the document that appeared a year ago.  If that appeals to you.  In the spreadsheet, please notice that in ConAgra’s Earnings press release (from 7/13/21)  the word inflation appeared 16 times (in the first 32.7 K characters of the 8-K, Item 9.01 document).  

For other document sections that you might want to use, see this list.

As a Calcbench user, you can certainly make this spreadsheet better or if you are really into working with text analytics, use our API to connect to the data repository yourself and write your own ( ahem… BETTER) version of the tool.  

Thanks again and enjoy the earnings season!

Wednesday, July 14, 2021

Pepsico ($PEP) filed its latest quarterly report this week, and we were skimming through the disclosures when something interesting caught our eye: the company listed a trio of acquisitions it had made in the last year.

That Pepsi makes acquisitions isn’t news; the company has anywhere from $15 billion to $20 billion in revenue every quarter and sits on $92 billion in total assets. Of course a firm that size can make plenty of acquisitions, including some plenty large ones.

In this latest report, however, Pepsi provided the purchase price allocation for all three deals — with quite a bit of variety among them.

First, let’s list the acquisitions themselves:

  • Pioneer Foods, a food and beverage company in South Africa, acquired in March 2020 for roughly $1.2 billion;
  • Be & Cheery, a large online snack company in China, acquired in June 2020 for roughly $700 million;
  • Rockstar, an energy drink manufacturer headquartered in New York, acquired in April 2020 for $3.85 billion in cash and another $880 million pending in contingent consideration, for a total price of $4.73 billion.

How did Pepsico allocate those deals? See Figure 1, below, for the disclosure details:

For Rockstar, the energy company, 48 percent of the deal’s total value was tied up in goodwill. For Pioneer Foods, 45 percent was assigned to goodwill. For Be & Cheery, it was 44 percent.

More interesting was that Rockstar’s deal had much less value assigned to physical stuff such as inventories and property, plant & equipment. That accounted for only 1.2 percent of Rockstar’s deal, versus 49.5 percent at Pioneer and 15 percent at Be & Cheery.

The other notable item here is that $2.4 billion paid to Rockstar for non-amortizable intangible assets. What’s that all about?

Honestly, we’re not quite sure. In theory, such assets would be any intangible asset with a perpetual life — say, distribution or franchise agreements with no expiration date. In Pepsi’s disclosures, we had a harder time figuring out exactly what that $2.4 billion is paying for. (Or the $309 million in the Be & Cheery deal, for that matter.)

The narrative disclosures under the purchase price allocation table only talk about goodwill value, and where each acquisition’s goodwill will be allocated among Pepsico’s various operating units. Then we jumped over to the firm’s disclosures about goodwill and intangible assets. That disclosure mentions “other indefinite-lived intangible assets,” which did increase from $17.61 billion in December 2020 to $17.74 billion as of June 2021 — but we still have no clear sense of what those assets actually are, and the year-over-year total increase doesn’t add up to what was reported in the acquisitions section anyway.

To be clear, we are not suggesting any shenanigans in Pepsico’s reporting. We merely point out that the details are important, and Calcbench has the detailed data that can let you, the analyst, ask better questions so you can get better answers.

We spoke with Don Pagach, Professor of Accounting and Director of Research for Enterprise Risk Management Initiative at North Carolina State University; Adjunct Professor at Duke University. 

Both North Carolina State University and Duke University are clients of Calcbench. Lucky us!

Tell us about your work as director of research for the Enterprise Risk Management initiative. 

The Enterprise Risk Management Initiative provides thought leadership around risk practices and their integration with strategy and corporate governance. As part of my research, I go into Business Risk Factors and look at Item 1A [of the 10-K]. For me, it’s interesting to see how the first business risk has changed over time. It’s also interesting to see how companies display information. Take Disney, for example. A little over a year ago, before the pandemic, Disney was focused on general economic conditions. Now, the first risk listed focuses on COVID. I like to see how companies display business risk information, what information is boilerplate, and what information is new. 

Another area I like to explore is how boards oversee risk. I spend a lot of time in proxy statements, looking at how companies handle risk. What I find easy with Calcbench is that I can get all the information on one screen. I can see the board oversight language and get insight into who’s responsible, and how risks are reported to the board. 

How do you use Calcbench for teaching? 

I teach both basic and more advanced accounting courses. In my Financial Statement Analysis course, we look at data to see how consistently companies report over time. I use Calcbench in the classroom to share timely and relevant examples, and to show all the financials in one place. What I like is that Calcbench shows what a company actually reports, not what is fed through a template. 

Students in the Financial Statement Analysis course use Calcbench for their final project,  to analyze financial statements of a company of their choice. With Calcbench, students are able to look at a company’s historical data. They are also able to compare performance to a set of competitors. If there are no direct competitors, students can create their own competitor set. 

In my Financial Accounting course, we go deeper into the footnotes to look at information such as non-controlling interest. The drill-down into financial statements is a nice feature. Students can go into the notes directly as opposed to scrolling through a 10-K document. 

How can we further help academics? 

The webinars and video tutorials are useful. There’s a lot of info in there. I would like to see more videos on how to use the various data sets.

Thursday, July 8, 2021

Manufacturing services business Jabil Inc. ($JBL) filed its latest quarterly report the other week, and we were intrigued to see that in the footnotes Jabil also discloses numerous key performance metrics to help investors see how tight of a ship management runs.

You can often find KPIs in the Management Discussion & Analysis section of the filing, and that’s where we found the numbers for Jabil. The company discloses five KPIs, as shown in Figure 1, below.

As you can see, underneath the KPI numbers Jabil also includes a description of how each metric is calculated. That’s in accordance with SEC rules, which say that a company can report a wide range of metrics so long as the firm (a) explains how the metric was produced; and (b) calculates and reports the metric consistently from quarter to quarter.

We call out Jabil’s reporting of KPIs just to remind everyone that Calcbench also lets you find a wide range of performance metrics even when a firm doesn’t report those KPIs directly — because the most important KPIs are based on specific GAAP disclosures, and Calcbench has those underlying data points, which means we can calculate various KPIs, liquidity ratios, and related metrics for you.

Finding the KPIs

You can find those numbers in several ways.

Our Multi-Company page allows you to search KPIs for one or more firms, by entering the name of the KPI in our standardized metrics field on the left side of your screen. For example, we searched “Days Payable Outstanding” for Jabil’s second quarter numbers, and received a result of 81.03. That matches the 81 days that Jabil reported for its quarter ending Feb. 28, 2021 almost exactly.

You can also then trace that KPI number to see how Calcbench arrived at it. (We always like to show our work around here.) Figure 2, below, shows how we found the 81.03 number for Jabil.

One point to note is that the exact wording that one firm uses to describe a KPI might differ from other firms, or from what Calcbench itself tracks. For example, Jabil uses “Days in Accounts Payable” for “Days Payable Outstanding,” but the formula to calculate them is the same.

So from time to time you may need to confirm that the KPI a firm reports in the footnotes is indeed the same KPI we calculate automatically. You can do that by double-checking what the firm says in the footnotes and comparing it to our number via the Trace feature. Investopedia is also a great resource if you want to see what goes into a KPI and other names it might be called.

You can also search for KPIs and other ratios using our Bulk Data page, which is sometimes the better choice if you want to research a large number of firms across long periods.

If you visit that page and scroll to the bottom, you’ll see a long list of KPIs that Calcbench can track. See Figure 3, below.

Just define the number of companies you want to search, across what periods of time, at the top of the page. Then check off the relevant KPIs listed in Figure 3 here, and hit Export. That’s all there is to it; you’ll get an Excel spreadsheet with the results in a few minutes.

By Jason Apollo Voss, CFA

Investors almost always have an opinion about the overall valuation level of financial markets. Differences in these opinions, of course, lead to different preferences for bids and asks — and hence, prices.

Markets are not necessarily rational; it’s just that the tug of war among investors disagreeing on valuations is more likely to lead to accurate pricing than not. One extreme view of markets is that prices are objectively determined, and an informationally efficient evaluation of the performance of businesses.

But what if these different opinions about the overall price of financial markets are based on a measure — the price-earnings ratio — that's susceptible to the subjective manipulation of businesses themselves, via financial engineering?

This leads me to ask the question: Are reported numbers really earnings, or just financial engineering?

Specifically, I’m talking about the ability of companies to control their total shares outstanding through issuance of more shares of stock and options and more importantly, via share buybacks. In short, chief financial officers have an easy ability to alter their firms’ shares outstanding. Which, to some extent, makes their earnings per share (EPS) a less than objective measure of performance.

It comes down to simple math. When you decrease the shares outstanding through a buyback, you magically manufacture higher EPS. CFOs can also probably create a higher valuation, assuming the P/E ratio increases or remains the same.

Therefore, investors need a more “financial engineering-free” P/E number to get a truer sense of valuation. This calculation turns out to be straightforward and is easy to calculate, as follows:

In other words, market capitalization divided by net income is our alternative measure. Yes, shares outstanding remains in the numerator. But it is completely absent in the denominator, which is where financial engineering has its biggest effect in manipulating the appearance of business performance.

Our P/E-Alt Measure in Action, Part I

What is the effect of financial engineering on a major index like the S&P 500? Here is what P/EAlt looks like for the S&P 500 since 2010. Row 1 is the traditional P/E measure, and Row 4 is P/E-Alt.

As you can see, for every year except 2010, the alternative P/E is much higher, meaning that the valuation is much richer for these businesses than initially appears to be the case. In other words, via financial engineering, companies are managing their EPS through share buybacks.

Typically this results in a pat on the back from investors, because companies make their earnings numbers. Firms also look more attractive as stocks because, on average, they appear to be less expensive.

On average, the difference in the two measures of P/E is 9.4 percent.  If we omit 2020’s outlier value of 39.8 percent, the average is still 6.4 percent. This effect is considerable and it helps to highlight two important things:

  1. The amount of manipulation of earnings per share. Said another way, business performance is consistently 6.4 percent worse than it appears.
  2. Just how rich valuations are for the S&P 500, because investors are willing to pay more for these companies than may be warranted by their actual profit performance.

At least we now have a more objective, more financial engineering-free measure of valuations.

Our P/E-Alt Measure in Action, Part II

But wait,  there’s more! Because companies can fudge their shares outstanding rather effortlessly, it also means that investors should not only look at the change in growth of EPS, but also at the change in growth of net income.

The difference between these two figures is also another way of examining the effect of financial engineering on a business.

Let’s use the same two companies I used in my last guest article for Calcbench where I compared Google ($GOOG) and Apple ($AAPL) to one another:

As you can see, Google’s net income has grown faster than its earnings per share: 16.8 percent compound annual growth rate in net income versus 16.1 percent for EPS. This means that Google issued shares at a rate faster than net income. We can also conclude because net income growth exceeds EPS growth, minimal financial engineering is taking place. Additionally, it’s encouraging to see the two figures so close to each other (16.8 percent and 16.1 percent), from a “we are not trying to fool you” standpoint.

In contrast, Apple’s EPS has a compound annual growth rate of 19.7 percent while its net income has a compound annual growth rate of 15.1 percent. In other words, its EPS has grown 30.4 percent faster than its net income. We may safely conclude that Apple is engaged in a much higher level of financial engineering than Google, based on the figures above.


When evaluating a company’s performance, it’s tempting to use earnings per share, or operating income per share, or EBITDA per share, and so on. But because companies have high levels of flexibility in controlling their total shares outstanding, investors need measures of performance that excise the effects of financial engineering as much as possible. In turn, this means we get a truer sense of performance and ultimately, valuation. Yes!

Editor’s note: If you want to conduct your own analysis along the lines of what Voss outlines here, Calcbench has created an Excel template you can download and use at home.

One detail: Voss accounts for stock splits manually, which are important in his examples. Our template allows the user to enter a stock split factor and have the calculation work for their case. If there is no split, the factor defaults to 1, making the template work.

We keep rolling today with our in-depth examination of SPACs and their related disclosure issues, with a look at one issue that’s made a lot of news lately: the wave of restatements that roiled SPACs earlier this year.

The broad strokes of the story are as follows. Back in April, the Securities and Exchange Commission published guidance warning SPACs (special purpose acquisition companies) that they may have been recording certain line-items incorrectly.

Specifically, the SEC said, many SPACS had been recording the warrants and FPAs they often issue to investors as equity, when the proper accounting treatment is to record those instruments as liabilities.

This is a rather complicated and arcane branch of accounting, so people shouldn’t assume that SPACs were doing this with nefarious intent. Still, the numbers involved are also large enough that for a firm to correct the reporting, it has to restate prior financials — which SPACs have been doing, in large numbers, for the last three months.

That wave of restatements also led to a big drop in the appetite among investors for investing in SPACs. According to an article in the Wall Street Journal recounting this whole predicament, the money raised by SPACs coming to market plummeted from $35 billion in March to only $3 billion in April, and monthly totals didn’t crack $4 billion May or June either.

Meanwhile, restatements shot up from 80 in all of 2020 to 330 so far this year. That’s the most since 2010, and this year still has six more months to go.

So what can Calcbench tell you about a SPAC’s restatements? Lots.

Looking for Restatement Data

Let’s go back to Pershing Square Tontine Holdings ($PSTH), the $4 billion SPAC led by billionaire investor Bill Ackman, which we first examined in Part II of our SPAC-tacular series. How would an analyst quickly find the particulars of Pershing’s restatement?

First, call up the SPAC in question using our Company-in-Detail page. You’ll see the company’s most recent financial statements, along with a series of tabs running above that display. Click on the option that says “Revisions.”

That will take you to a table that lists the SPAC’s recent filings. Any filings that have been revised will include a small + icon next to the relevant period. See Figure 1, below. The important parts are noted in blue.

Right away we can see that on May 24 Pershing restated its 10-K filing from earlier this year, and that the restatement included 14 adjustments.

Exactly what were those adjustments? Click on the + icon to the left there. That will expand the page to show each revision made by amount, line item, and date. See Figure 2, below.

Now we’re getting to the good stuff. For example, we can see that liabilities ballooned from $57.8 million in the original 10-K filing, to $1.114 billion in the restated version. That makes sense; the SEC’s notice had told firms to report all those warrants as liabilities, after all. (Note that in the Revisions column you can see the exact date of the original filing and the subsequent restatement, too.)

The restatement also increased Pershing’s net loss from $1.4 million to $954.8 million; and cut shareholder equity, too. Again, one would expect those changes given the nature of the accounting issue that Pershing had to restate.

Some SPACs have had more extensive restatements than others. Apollo Strategic Growth Capital ($APSG), which reported $818 million in total assets for Q1 2021, made a boatload of revisions across numerous periods this year— including 39 revisions made to its 10-K. See Figure 3, below:

The above are only a fraction of the revisions Apollo made to its 10-K. And they’re not to be confused with the other revisions Apollo also made to its 10-Q filing from mid-June, either.

Calcbench can’t speculate on what this restatement wave means for SPACs over the long term. Perhaps it’s only a temporary pause while everyone adjusts to the new numbers on the balance sheet; maybe other pressures that have nothing to do with warrants and GAAP accounting rules will quell SPAC mania anyway.

Our point is only that if you need to see exactly what a SPAC has been doing, or revising, we have all the data you need in just a few keystrokes.

Today our series on SPACs (special purpose acquisition companies) looks at perhaps the most important question of all. Where do they spend the gobs of money they raise?

After all, remember the natural lifecycle of SPACs. They go public first, raising cash from investors in an IPO; and acquire an operating business later, where both entities are then reborn as a publicly traded operating firm.

Moreover, SPACs generally have 18 to 24 months to find their acquisition target and “de-SPAC” into that merged operating business. If the SPAC can’t close a deal within that window, it must (you may want to start breathing into a paper bag here) give the money back to investors.

So we wondered: how can one find those IPO dates and merger timelines? And how can we find details about any de-SPAC mergers that have happened so far?

IPO Dates and Deal Deadlines

Let’s start with finding IPO dates and time horizons to close deals. As an example, we’ll use Khosla Ventures Acquisition Corp. ($KVSA), a SPAC ultimately managed by venture capital titan Vinod Khosla, co-founder of Sun Microsystems and head of Khosla Ventures.

First, find the filings for Khosla Ventures Acquisition in the Calcbench Interactive Disclosure database, and specifically pull up the Description of Business Operations disclosures. (That will always be the place to start when researching any SPAC.)

There, we can see that the Khosla SPAC was founded on Jan. 15, 2021. Four paragraphs down, one can also find that Khosla went public on March 8. That IPO generated proceeds of $345 million.

Next, we wanted to know how long Khosla had to put that money to work. So we just did a simple text search for “months.” Sure enough, several paragraphs further down we found this:

If the company is unable to complete a business combination within 24 months from the closing of the Initial Public Offering (the “Combination Period”), the company will (i) cease all operations except for the purpose of winding up; (ii) as promptly as reasonably possible but not more than ten business days thereafter, redeem the Public Shares...

That’s our answer. Khosla has until March 8, 2023 to consummate a merger with an operating company out there somewhere. When researching your own SPACs, you may also want to search for “years” or “18”, “24”, “two years” or similar terms. That should bring you to the detail you want to find.

Acquisition Disclosures

OK, so the clock is ticking for Khosla to close a deal. Then we noticed that in addition to its Q1 2021 filings, Khosla also filed an 8-K on June 9. Had these folks already found a target?


The 8-K filing announced that Khosla Ventures Acquisition had struck a merger deal with Valo Health, a drug discovery startup in Boston using artificial intelligence to do… um, something with pharmaceutical research. We’re just financial data geeks here.

The deal values Valo at $2.8 billion, and is expected to close sometime in the third quarter. It also involves numerous other firms investing $168.5 million in the resulting public company in a PIPE (private investment in public equity) deal, which is not uncommon to see in SPAC land.

What we don’t yet know is the purchase price allocation of this deal, since (a) it hasn’t actually happened yet; and (b) firms typically don’t disclose purchase price allocation details until several months after deal closing anyway. So we don’t know how much of Khosla Ventures’ cash will go to goodwill, tangible assets, and so forth.

Finding Details on Purchase Price Allocation

Hold up, though. We don’t know purchase price allocation here because this de-SPAC transaction is still pending. What about prior deals from other SPACs? Can one find out those numbers?

In Calcbench, yes you can.

Start by going to our Segments, Rollforwards & Breakouts database. There, you can select the “Business Combinations - Purchase Price Allocations” option from the drop-down menu at left; and from the “Choose Companies” option at the top you can screen for all firms that use SIC code 6770 — the code for SPACs. To get the maximum amount of data possible, we also checked the All History button that’s on the top right.

Figure 1, below, shows what we did and some of the results. Note the history check-box and the disclosure choice, both circled in red; and our SIC 6770 screen, flagged with a blue arrow.

The results are the purchase price allocation by type, for each transaction. That’s why the first three results all look similar; they all pertain to an acquisition completed by UPD Holding Corp. ($UPDC) in the first quarter of 2021: cash, right of use assets, and goodwill.

So UPD apparently spent $480,052 to acquire a business earlier this year, and nearly $417,000 of that sum went to goodwill. To learn the exact details of this transaction, you can then hold your cursor over one of those line items and use the world-renown Calcbench trace feature to go to the full disclosure. Which we did, and we found this:

In February of this year, UPD agreed to acquire a drug treatment business in Kentucky, Vital Behavioral Health Inc. That business consists of one outpatient treatment clinic, and one residential facility for people in addiction recovery. On May 21, UPD detailed the purchase price allocation. Almost all of the $480,000 price went to goodwill, with a smattering paid in cash and right-of-use assets.

Read the fine print, however. You can see in the narrative disclosure that neither of Vital’s two facilities had any licenses to operate as of May 21. Hmmm. Moreover, if you then jump over to the business description for UPD, you can find that the business previously operated as a brewing company, Record Street Brewing, which UPD sold at the end of 2020 before getting into the addiction recovery business.

We’ll say no more about the merits of the UPD-Vital deal. Just understand that the data is out there, and Calcbench can help you find it.

Today we continue our series on SPACs (special purpose acquisition companies) by examining what these firms actually report on the balance sheet and the income statement.

Our first post in this series examined SPACs at the global level: how many such firms actually exist (at least 350 in first-quarter 2021) and the total assets all these firms have (roughly $109 billion, a ten-fold increase from early 2020). We found eight firms reporting more than $1 billion in assets, and only one of them had more than $2 billion: Pershing Square Tontine Holdings (PSTH), managed by billionaire investor Bill Ackman.

So let’s start there. What does a SPAC like Pershing report?

On the income statement Pershing doesn’t report too much, because so far the SPAC isn’t doing much. The company had $4.78 million in legal fees and another half-million or so in expenses in Q1, and a piddling $638 in interest income. That means an operating loss of $5.3 million for the quarter. See Figure 1, below.

But wait, you say! What are those two line-items for change of fair value in liabilities? Together they add up to $341.6 million in other income, which drove an impressive $337 million in net income for the quarter.

The warrants and forward purchasing agreements are other financial instruments that Pershing (or any other SPAC) offers to investors alongside its IPO shares. They are recorded as liabilities on the balance sheet. The SPAC must then assess the fair value of those liabilities every quarter.

In Pershing’s case, the fair value of those warrant and FPA liabilities declined, which is then recorded as Other Income on the income statement. See Figure 2, below.

You might also be wondering, “Is this warrants and liabilities stuff the same issue the SEC flagged about SPACs in April, and caused a wave of restatements?”

Yes, it is. The SEC warned SPACs that they had been recording warrants and FPAs as equity, when the proper accounting treatment was to record them as liabilities. After that warning many SPACs, Pershing included, had to restate their prior financials. We’ll take a deep dive into that issue in a subsequent post. Figure 2, above, shows how warrants and FPAs are supposed to be reported. (Pershing had fixed the matter for its Q1 report.)

As a practical matter, however, Pershing’s Q1 net income is almost entirely notional. The SPAC did not actually rake in hundreds of millions in revenue. The operating part of its income statement is pretty much a snoozefest.

And the Balance Sheet

The balance sheet isn’t much more exciting. It reveals that Pershing has several classes of assets such as cash, prepaid expenses, and dividends — but the amounts haven’t fluctuated all that much over the last few filing periods.

The big item on the balance sheet is the “cash and securities held in trust account.” That’s where a SPAC parks the capital it raises from investors, before spending that cash on an acquisition target. As you can see in Figure 3, below, Pershing has $4 billion at its disposal.

In fact, the only significant change on the balance sheet has been those FPAs and warrants, and the change in fair value Pershing recorded in Q4 2020 and Q1 2021. Most other items hold fairly steady.

As always, you can dive into this data deeply with Calcbench. For example…

U.S.-listed firms are required to follow U.S. Generally Accepted Accounting Principles (GAAP) when reporting their financial statements. Firms are also allowed to present additional figures that don’t follow GAAP — metrics commonly known as “non-GAAP.” 

When doing so, the firm must also explain how it arrived at the non-GAAP figure, and show a reconciliation (including the adjustments made) back to the closest comparable GAAP metric. 

In a recent Calcbench research report, we examined the size and nature of adjustments S&P 500 companies made to their GAAP figures, to arrive at the most commonly used non-GAAP figures. Today we want to examine the trends of those differences. 

First, we identified all companies that report non-GAAP earnings per share (EPS), roughly 1,700 firms. Then we examined the median difference between the non-GAAP EPS and the corresponding GAAP EPS. 

One interesting note is that the percentage of companies reporting non-GAAP EPS has increased significantly, from about 20 percent of all firms in 2013 to 30 percent in 2020. This would suggest that investors find the non-GAAP information important, and push companies to disclose it. 

Meanwhile, the median difference between non-GAAP and GAAP EPS has doubled, from roughly $0.30 per share in 2013 to $0.60 in 2020. The non-GAAP EPS is higher (little surprise there), and the gap between the non-GAAP EPS and GAAP is increasing


Does this mean GAAP numbers are less important, or that non-GAAP amounts are more informative? We don’t know, but the increasing gap is interesting.

We further examined the distribution of differences between GAAP and non-GAAP EPS for 2020. As you can see from the table below, 80 percent of companies report a non-GAAP EPS that is higher than the GAAP EPS, whereas only 20 percent report a non-GAAP EPS lower than the GAAP EPS.

We wanted to further examine the relationship between the GAAP EPS and non-GAAP EPS, so we plotted the EPS numbers for 2020 in a scatter plot, shown below.

As you can see from the chart, GAAP and non-GAAP EPS largely move in similar directions. Still, you also see some cases where the two seem to differ substantially. There seem to be cases in the above chart where the dot falls below the 45-degree line. Those would be cases where non-GAAP EPS is lower than GAAP EPS.

Then we zoomed in to focus on EPS values of $10 or less. In the chart below you see a similar story to the one we saw for the entire population, with the exception that there seem to be many more cases above the 45-degree line — that is, where non-GAAP EPS is higher than GAAP EPS.

Want more? Calcbench has all the data. Go to our Multi-Company database page to see the data. (Caveat: non-GAAP data is available for Professional-level Calcbench subscriptions only).  

Saturday, June 12, 2021

Like so many others who study financial data and follow financial news, Calcbench has been fascinated lately by special purpose acquisition companies — more commonly known as SPACs.

What are these financial reporting contraptions? How many are there? What financial disclosures do they make, and how can one study those disclosures for useful financial analysis? How much do the people behind SPACs get paid?

Calcbench decided, therefore, to launch a short series on this blog looking at financial reporting issues related to SPACs. Because these firms do indeed submit financial data, which means we do indeed have that data available for our subscribers to find and put to use.

This first post seeks to answer some basic questions about SPACs. In subsequent posts we’ll examine specific SPAC issues in more detail.

What are SPACs, exactly?

SPACs are publicly traded holding companies, and more commonly known as “blank check companies.” They have no substantive operations; they exist to hold money in trust from investors, and then go hunting for a private operating company to acquire.

Once that acquisition is done — presto! The newly merged entity is a publicly traded operating company.

In theory, SPACs are a faster, simpler way for private companies to go public than a traditional IPO. Of course, SPACs also make gobs of money for the executives who put the SPAC and ensuing acquisition together (known as “sponsors” of the SPAC) and often for the entertainers, athletes, and other famous people who serve as “advisers” to the SPACs.

How many SPACs exist?

That’s a good question, because a legion of firms have announced themselves as SPACs over the last 18 months or so.

One good place to begin is with SIC codes. SPACs are category 6770. When we set up a peer group for all firms in that SIC category as of Dec. 31, 2020, we found 189 “live” firms that had submitted current financial statements — everyone from 5:01 Acquisition Corp. ($FVAM) to Zeta Acquisition Corp. I ($ZETAI).

Jump forward to Q1 2021, however, and the picture looks markedly different. We found 358 firms with current financial statements, including new entrants such as 26 Capital Acquisition Corp. ($ADER) and Pershing Square Totine Holdings ($PSTH), billionaire investor Bill Ackman’s SPAC.

That said, it’s worth remembering that plenty of firms announce that they’re forming a SPAC, and even file an S-1 registration statement with the Securities & Exchange Commission. Then… nothing. They ghost the capital markets, never to file a statement again. So when we say “189 blank check companies in Q4 2020” or “358 firms in Q1 2021,” we’re talking about firms that have filed quarterly statements for that period and are keeping up with their disclosure duties.

How much money do these SPACs have?

Lots, and most of it raised only within the last year or so. We used our Bulk Data Query tool to track total assets and average assets for all firms under SIC code 6770, from the start of 2019 through first-quarter 2021. Figure 1, below, tells the tale.

Huge amounts of money poured into SPACs in the first quarter of 2021, when the SPAC frenzy seemed to be at its height. Total assets soared from $49.1 billion in Q4 2020 to more than $109 billion in Q1 2021, while average assets per firm jumped from $231 million to $311 million in the same period.

We also saw the number of large SPAC firms (defined as having $1 billion or more in assets) rise from three at the end of 2020 to eight at the end of Q1 2021. Here are the largest 10 firms at end of the quarter:

And what do SPACS do?

As we said earlier, they raise funds from investors first, and then go acquire an operating company later. Many SPACs have a specific theme — say, only biotech companies, or only European businesses, or only businesses owned by women or developing ESG products or whatever.

The crucial element is that SPACs have two years to find their acquisition target; if they haven’t closed a deal by then, the SPAC managers must — horror of horrors — give the money back to the investors.

Which creates all sorts of pressures to close deals quickly. Which we will explore in our subsequent SPAC-tacular series of posts.

At Calcbench we love seeing how people use our data. Particularly rewarding, given our many years spent in the classroom, are the citations by professors conducting academic research. 

A couple of years ago, we shared several articles written by our academic friends. Given the success of this post, we thought we’d update our blog readers with some of the latest research. Note: to see other topics that utilize Calcbench data, head to and type “Calcbench” in the search bar. You’ll find nearly 100 results. 

Below are some recent examples of papers that include Calcbench mentions: 

For some additional mentions visit our academics page. Dozens of universities are using Calcbench today to access reliable financial data and information deep within financial statements for their research needs. 

(Editor’s note: Today we’re delighted to introduce a new monthly column on the Calcbench blog that will explore financial analysis issues, written by Jason Apollo Voss — investment manager, financial analyst, and these days CEO of Deception and Truth Analysis, a financial analytics firm. Enjoy, and be sure to see our companion spreadsheet to help you explore the concepts Voss talks about.)

By Jason Apollo Voss, CFA

In today’s world where the biggest assets on the balance sheets of the most successful firms are all about intangible assets, you may wonder why people should still care about understanding property, plant, & equipment (PP&E) better and more deeply. The answer is twofold:

  • Many firms play games with depreciation & amortization to improve their short-term results, and such games depend on PP&E.
  • Even the biggest tech firms have big investments in hard assets.

In fact, from the details of PP&E disclosures you can draw numerous conclusions about growth capex and maintenance capex, two concepts crucial to financial analysts. 

So let’s consider the two above points in turn, using Google (GOOG) and Apple (AAPL) as a compare-and-contrast case study.

Big Tech Firms’ Big Investment in Fixed Assets

Google reported 2020 total assets of $319.62 billion. That’s certainly impressive, and most of those assets must be intangibles such as goodwill, right?

Nope. Google’s total intangible assets, including goodwill, are just $22.62 billion — 7.1 percent of total assets. See Figure 1, below. Hmmm. 

In contrast, gross property, plant, & equipment for Google is a whopping $126.46 billion, or 39.6 percent of total assets. In other words, Google’s fixed assets dwarf its intangible assets by a factor of 5.6. (Net PP&E, shown above, is “only” $84.749 billion.)

What about Apple? It outsources so much of its manufacturing, and the company is all about intellectual property. So surely its IP must dwarf its fixed assets, right?

Apple’s total assets for 2020 were $323.89 billion. Its IP assets were… um, hold on. Apple doesn’t actually report its intellectual property in detail, either on the balance sheet or in the footnotes. What’s that about?

First, a small digression. Apple’s intellectual property has to be its most important asset, and yet there’s almost no detail about it in the 10-K. As an investor, this kind of fog worries me. Essentially the company is saying, “Trust us, we are Apple.”

Back to the numbers. Apple does report an aggregate category named “Other Non-Current Assets.” In 2020 those assets totaled $42.52 billion, or 13.1 percent of total assets.

What bothers me is that things like deferred tax assets are also embedded in that number. So investors in Apple, likely enamored by its intellectual property, can’t actually get a direct read on the value of that IP.

Apple does disclose in its footnotes the size of its net deferred tax assets, which totaled $18.30 billion in 2020.

With this number, buried in the footnotes, we can then estimate that IP number indirectly. It’s likely to be Other Non-Current Assets ($42.52 billion) minus net deferred tax assets ($18.3 billion), which leaves us $24.22 billion for estimated intellectual property intangible assets, which is about 7.5 percent of total assets.

And what about Apple’s fixed assets? Those mundane Dickensian assets, the ones that surely don’t matter to investors?

In 2020 those assets totaled $103.53 billion — roughly 32 percent of total assets, and 4.3 times larger than estimated intellectual property intangible assets.

I think you see my point: the stories of Google and Apple ought to be enough to convince you as an investor that you should pay attention to the goings on within fixed assets, even in tech companies. That is, big tech needs BIGGER MECH.

PP&E Analysis Nuances

Because most firms provide only a single line item on their balance sheets about property, plant, & equipment, it’s difficult to determine whether the firm is keeping up with maintenance capital expenditures.

This is because depreciation is ultimately a cash expense, even though it’s typically treated as a non-cash expenditure. After all, you can only drive your car for so long without investing in repairs and maintenance before it no longer works. True, day in and day out the depreciation appears to be a non-cash expense — until the day that it isn’t. Ouch.

If an analyst is building a cash flow valuation model, he or she typically has a conversation with someone from the company in which they invest and asks for an estimate of maintenance capex. But we usually can estimate maintenance capex directly from the information reported in the annual report. We can also see how companies are aging their assets. Here’s how.

There are several PP&E metrics that you should know and incorporate into your fundamental analyses. Why? They lead to a more nuanced (read: higher probability of generating alpha) understanding of business performance. These metrics are:

  1. Average expected life of fixed assets
  2. Fraction of expected life exhausted
  3. Average remaining years =

(1-fraction of expected life exhausted × average expected life of fixed assets)

With these numbers you can do a magical thing: gain insight into how companies under-invest in fixed assets in the short run, to create the appearance of increased earnings and increased operating cash flow; and insight into underreported cash flows from/used in investing.

Average Expected Life of Fixed Assets Calculation

Here’s our first calculation, to determine average expected life of fixed assets. 

If you think about this formula it makes logical sense. Namely, if we want an estimate of how many years the fixed assets have left, we should take the total depreciable fixed assets and divide it by the amount taken most recently for depreciation and amortization. This gives us an idea of the average number of years of expected life.

Google, like most firms, has a single line item on its balance sheet labeled “Property & equipment, net.” But the net figure is distorted by accumulated depreciation, so we need the gross figure.

The numbers for gross PP&E and for accumulated depreciation are sometimes located on the balance sheet itself, or almost always in a footnote. In Google’s case for 2020, those amounts are found in “Footnote 7. Supplemental Financial Statement Information: Property and Equipment.”

In any case, the amount for Google in 2020 is $126.462 billion. But land is not a depreciable asset, so we need to subtract its value out of the gross property, plant, & equipment figure or else it will make the Average Expected Life of Fixed Assets figure too high. Again, the total value of land for Google in 2020 was reported as $49.732 billion.

Next, we need the depreciation & amortization expense figure for a single year. It’s located typically in one of the following places: the income statement (rarely), in a footnote somewhere (rarely), or on the cash flow statement in the operating cash flow category (almost always). For Google in 2020, that number is $12.905 billion.

With our figures in hand we can now calculate the average expected life of fixed assets for Google as:

Before considering our next nuanced PP&E measure, we need to do a qualitative check. Does 5.95 years for Google’s depreciable fixed assets make sense? What are those assets exactly? Most likely, they’re office buildings owned by Google, as well as a massive investment in things like server farms.

One way to answer this question is to look at the composition of depreciable fixed assets in Google’s Footnote 7, and compare it to the company’s accounting policies regarding depreciation. We could even calculate a weighted average expected depreciation for Google. That’s outside the scope of this article, but it’s possible.

Fraction of Expected Life Exhausted

Here is how to calculate our next nuanced metric. Fraction of expected life exhausted equals:

Again, this formula makes logical sense because it looks at the total accumulated depreciation over the years that the firm has owned its depreciable fixed assets; and then compares that number to the original purchase/book value of those assets. This is a reasonable estimate for how much of the depreciable fixed assets’ life has been exhausted.

For Google, this is calculated for 2020 as follows:

$41.713 billion ÷ $76.730 billion = 54.4 percent

In other words, while Google has booked assets that have an average expected life of 5.95 years, 54.4 percent of the value of those assets had been depleted by the end of 2020.

Average Remaining Years

Our next metric to calculate is average remaining years. It can be calculated as:

(1- fraction of expected life exhausted) x average life of fixed assets

In the case of Google, we can estimate that number as follows:

(1 - 54.4%) x 5.95 years = 2.71 years

The small number of years tells you something important. Namely, Google cannot afford to skimp on maintenance capex for too long before problems would likely start to emerge with key assets like servers.

Now, if we only had a way of estimating maintenance capex from our data…

Maintenance Capex

While the above figures were not created by me, to my knowledge the following maintenance capex estimate is my own creation. (Surely others have seen in the data a chance to estimate this key figure in valuation models.) Here’s the calculation:

Maintenance capex =

The innovation here is really not so tricky. If we’re trying to estimate this year’s maintenance capex spend, we cannot take this year’s depreciation & amortization expense because we need the total depreciable fixed assets from the beginning of the year. Consequently, in the denominator of the above equation, we add it back to the total accumulated depreciation for the firm.

For Google, it looks like this:

Voilà! But how does this compare to the actual capital expenditures (that is, purchases of property and equipment) taken by Google in 2020?

Well, Google reported $22.281 billion on its cash flow statement in its Investing Activities sub-section. If you are doing a valuation model for Google, and if you had calculated the above metrics in a time series, then you could independently (from management’s input) estimate the proportion of all capex that ought to be spent each year by averaging them together.

For the single year 2020 for Google, that figure is 81.4 percent. Wow! In other words, Google rapidly depletes its assets. This makes sense given that it is a tech company.

With the above figures you can also look at an interesting time series of over- or under-investment. If a company under-invests in maintenance capex, based on your estimate, it begs several questions:

  • How sustainable is the underinvestment? Will the company’s competitiveness be hampered if it chronically underinvests?
  • Why is it underinvesting in capex? Is management trying to deceive me into believing that the firm is performing better than it actually is?

But we can now infer another interesting conclusion from the above estimate for maintenance capex. Do you see it hidden in there?

We also now have an estimate for Google’s growth capex.

Growth capex can become the denominator in an interesting ratio (again, I think of my own invention):

Basically, this ratio looks at the marginal revenues generated by the growth capex spent a number of years ago. How many years? A good estimate for the number of years is the average of the time series of the average expected life of depreciable assets figures.

Next, as investors we should expect the maintenance capex figure to be lower than the total capex figure. But that isn’t always true. Let’s bring Apple back into the picture and compare it side by side with Google. Buckle up!

(Yes, the below image is a bit hard to read; Calcbench also has an Excel spreadsheet with the same formulas built into the model available for download, if you’d like to see other examples.)

At a glance we can see several interesting things going on. In particular, note that Apple’s time series for these figures is more volatile than that of Google’s.

For example, Apple’s coefficient of variation (that is, standard deviation  average) for its Average Expected Life figure is 24 percent, versus just 6 percent for Google. As investors, this is an indirect way to see that Apple’s business is likely more complex and slightly more volatile than that of Google’s.

Also noteworthy, and confusing, is that our estimate of maintenance capex for AAPL exceeds its actual capex. And we’ve been generous in these figures, because we are adding to total capex the value of Apple’s acquisitions. But Apple reports this figure “net,” meaning that its sale of businesses may be financing its capex — although we can’t say that for sure. Still, this is another example of not being able to decipher the full picture of what Apple is doing. Hmmm.

In my book, co-authored with C. Thomas Howard, Return of the Active Manager, I wrote about the behavioral insights that are embedded in financial statements. A hypothesis, I think, can be safely inferred from our deeper dive into Property, Plant, & Equipment Nuances. Namely, Apple needs higher scrutiny than Google, and Apple seems to behave in an obfuscating way.

Thursday, May 27, 2021

The NYT published an opinion piece this morning, The Apple Tax is Rotten by Farhad Manjoo.  The piece focuses on the recent court case (judgment pending) regarding the App Store.  We wanted to learn more.  Please keep in mind that our goal is not to critique the piece or the author.  We thought we ought to see how quickly we could inform ourselves, and our users about what could be learned about financial disclosures and particularly segments from Apple Inc’s latest filings, especially because that is what we do here at Calcbench.  

So, we dug right in.  

We went to the Disclosures section of Calcbench and looked up Apple’s segment disclosure to learn more.  It looks like this:

Note that there is nothing about the App Store here.  The only thing listed are geographies.  So, we continued and found that to obtain product level revenues, we need to go to the Revenue Recognition disclosure, where Apple breaks out the product revenue. 

Once you get there ( below picture), you can either download the data directly, or use the Calcbench Excel Add in to build some formulas and create a time-series.  

Here’s the breakdown of product revenues in dollars as well as percent of total.  Please note that the Apple Inc fiscal year ends around September 30 so the last bar in our time series graph represents the 6-month period ending on March 31, 2021. 

Quick conclusion, the iPhone is still king, but the services revenue is significant and growing in dollar terms as well as a piece of the overall pie (in $ B) .  

Wednesday, May 26, 2021

Calcbench just published our latest quarterly analysis, for firms’ collective financial performance in Q1 2021 — and folks, we need to talk about net income. It went nuts.

We examined the Q1 financial disclosures of more than 3,400 non-financial firms, comparing this year’s numbers to first-quarter 2020. Net income rose by more than 1,900 percent, from a piddling $16.2 billion one year ago to more than $330 billion today.

In the abstract sense, that rebound isn’t a surprise. Q1 2020 was hammered by the COVID-19 pandemic, so naturally Q1 2021 would be much better as vaccinations spread across the United States and the economy kept reopening.

In the practical sense, however, with actual numbers staring you in the face — wow. We don’t even know what the correct verb should be here. Net income soared? Skyrocketed? Spiked? You tell us.

The rest of our Q1 2021 wrap-up (you can download the full analysis from our Research Page) tells a similar, if less dramatic, story. Revenue, cash, and cashflow from operations all rose at respectable rates; debt only crept up 1.3 percent; and operating expenses declined by 12.2 percent. Presumably that’s because firms were spending less on emergency supplies such as PPE or cleaning equipment.

Figure 1, below, shows the year-over-year change for all 3,400 firms we studied, across 11 types of financial disclosure. (Net income not included, because its 1,944 percent spike would ruin the visual representation of everything else.)

Beyond this aggregate analysis, we also studied a smaller group of roughly 1,500 firms in across 20 industry categories classified by SIC code: pharmaceuticals, surgical & biological goods, oil & gas, retail, motor vehicle parts, and others. Then we compared each industry’s quarterly numbers across seven major financial disclosures.

Figure 2, below, shows an example for cashflow from operations.

So enjoy our latest quarterly wrap-up, and check back in three months or so for Q2!

Tuesday, May 25, 2021

Lots of people assume that accounting changes are boring and unworthy of one’s attention. Not true, we say! On many occasions, those changes can have a significant impact on the financial statements.

Take Pfizer (PFE) as one example (mainly because the company is so high-profile thanks to its coronavirus vaccine).

Pfizer filed its latest 10-Q report on May 13. Tucked away in the footnotes, the company disclosed that it has adopted a new accounting principle for the valuation of its pension plans. As described in Note 1(c) of the filing:

In the first quarter of 2021, we adopted a change in accounting principle to a more preferable policy under U.S. GAAP to immediately recognize actuarial gains and losses arising from the remeasurement of our pension and postretirement plans (“MTM Accounting”).

As a result of that change, the amount reported on Pfizer’s balance sheet in the 2020 10-K was revised in Q1 2021.

The important part: you can see such changes in Calcbench, since we highlight these revisions. See below:

From an accounting perspective, what happened was this: instead of deferring accumulated losses in Accumulated Other Comprehensive Income (AOCI), Pfizer recognized those losses now and decreased Retained Earnings for the corresponding amount. Therefore the amount for Accumulated Other Comprehensive Income (AOCI) was revised from an $11.7 billion loss to a $5.1 billion loss.

Let’s say that again more simply: the accounting change led to a $6 billion increase in AOCI and a corresponding decrease in retained earnings. All because Pifzer adopted a change in accounting rules.

Now that the pension and post-retirement plans are “marked to market,” their carrying amount would be subject to market fluctuations. That can introduce some volatility in the future.

Interesting side note: Pfizer released its earnings report and filed an 8-K on May 4 — but those filings included no mention of the change in accounting principle and revised numbers. Analysts had to wait for the subsequent 10-Q filing and, as always, look for the footnotes!

Thursday, May 20, 2021

Another quarter, another expansion of Calcbench’s awesome financial data superpowers. This time around, we’re here to tout updates to our earnings press release tools.

The news is this: that Calcbench Professional users can now access all numbers in the text and tables of firms’ earnings releases, within minutes of said press releases hitting the wires. In a recent study we conducted of more than 13,000 press releases, our users had access to data within press releases in an average of five minutes.

Those enhancements enable financial analysts and money managers to make more informed decisions, using Calcbench’s model-ready data. As with all Calcbench data, earnings press release data is available to export from our website, or can be directly accessed through Excel, Google or the Calcbench API. (You can also see earnings releases on our Recent Filings page. It will look like Figure 1, below.)

One recent example is Iron Mountain ($IRM), an enterprise information management company. On May 6, prior to 7:00 a.m. ET, Iron Mountain released its earnings press release. By the time the market opened, Iron Mountain’s stock price gained $0.78. By the closing bell, Iron Mountain’s price increased another $2.14. Analysts who had systemic access to the press release data would have had their models primed to take advantage of those price moves.

“Information moves markets. Calcbench understands that the more efficiently analysts can populate their models, the faster they can act on the new information.” said co-founder and CEO of Calcbench. Pranav Ghai. “That’s why immediate access to data embedded in earnings press releases, such as GAAP, non-GAAP, key performance indicators and segments, is critical.”

We agree, and not just because Ghai is our boss. The point is valid no matter who says it — and because it’s valid, that’s why we’ve developed that access to the data for our users.

By the way, Calcbench continuously upgrades its functionality based on user feedback. We previously announced the ability to compare side-by-side comparisons of preliminary income statements against previously reported numbers, without the hassle of manually inputting the data.

What will we do next? Stick around and find out.

FREE Calcbench Premium
Two Week Trial

Research Financial & Accounting Data Like Never Before. More features and try our Excel add-in. Sign up now to try the Premium Suite.