Determining Market Capitalization With Extreme Consistency And Accuracy

Summary

• Short Levi Strauss. Its longer-term valuation is between $9.82 and $10.11.
• Buy and hold Intel. It has the potential to reach a revenue $118B from the current $77B.
• Sell Qualcomm as it is the most leveraged by far in its sector.
• This is a powerful tool for IPO, M&A, Angel, VC managers and investors to value share price before and after IPOs.

Introduction to my Background and Why You Should Read This Article

Everything in the pricing of risk can be traced back to (some version of) the Dividend Discount Model (DDM). This is an analytical model. Like all analytical models the DDM has explicit and implicit assumptions (also known as axioms) built into these models. It is the implicit assumptions that worry me. For example, one of DDM's implicit assumptions is that the market cares about the cost of funding. Nobody I know has tested this. In this article I show using numerical modeling, that many times the market does not care about the cost of funding. Shocking, isn't it. That is, cost of funding is a cash flow problem, not an economic problem.

When you read this article, have an open mind, and put aside everything you think you know about financial risk. Empty your cup, as I am proposing a rigorously tested quantifiable new approach to financial risk, based on now my more than 35 years of analytics in many different fields.

Let me give you a brief background of my research endeavors. Most of you will know me from my two Seeking Alpha articles "Is The Recession Here?" and "The Fed's Bank Stress Test is Wrong."

There are two parts to my background. First, having successfully completed my Masters in Operations Research in 1982, with Lancaster University, UK, I completed my Master's in Finance in 1995 with the University College Dublin, Ireland. This article is written using years of experience in both these technical fields.

Since I had proved that a portfolio's unsystematic risk cannot be fully diversified away, I reworked all our definitions of risk. My shortened 553-page finance thesis titled Unsystematic Risk is available on Amazon. This thesis is based on 2,000 years of daily real and simulated stock prices from 15 exchanges and indices around the world. That is the kind of detailed perseverance it takes to find real and rigorous quantifiable answers in finance - the pricing of risk as opposed to accounting which is about the fidelity of the book value of assets and liabilities.

For example, it is possible to calculate the unsystematic risk present in a portfolio and real-world experience shows that if a major company (e.g., GM, Enron, etc.) collapses, that will take down your portfolios. I had quantified the Efficient Market Hypothesis. Proved that betas are short term (less than 90 days) metrics. I then introduced numerous tools to manage portfolios.

The second part. Based on my analytical/numerical modeling skills, these past 20 years, I have rewritten the foundations of physics (the link to the list of my peer reviewed publications and books are here). In 2007 I discovered an elegantly simple formula for gravitational acceleration g=τc², that does not require we know the mass of the planet or star which proved that all string theories are wrong. This formula is considered sacrilege in theoretical physics because there is no mass. There is much, much more, but this is sufficient for this article.

Investors Consists of Short- and Long-Term Players Focused on Different Types of Risk

We are interested in both the downside and the upside risks of a company. I have addressed downside risk in my article "The Fed's Bank Stress is Wrong," and therefore, will not address it here. Bear with me as I have changed our perspective on financial risk and some of these definitions. So, empty your cup.

In a securitized market like the stock exchange, a share is a unit representation, not of ownership - though that counts - but of the underlying corporate Economic Engine. This Economic Engine is chockablock full of unsystematic risk. Can we, from financial statements, get a handle on the unsystematic risk? Yes. I derive the Economic Statements from Financial Statements, to provide a robust handle of this unsystematic risk. See Fig. 1, Partitioning Sources of Risk. Moving from left-hand side to right-hand side to better understand market mechanism.

As shown on the right-hand side of Fig. 1, Market Risk is composed of investors' Expectation Risk and Corporate Risk. If you think that unsystematic risk is fully diversified away in portfolios, think again. Ask yourselves what happened to your portfolios when Old GM and Enron collapsed.

Fig. 1: Partitioning Sources of Risk (Source: Ben Solomon, 2019)

What do we do with this knowledge? We can determine with a very high degree of accuracy the median share price. Any price fluctuations about this median is Expectation Risk i.e. some form of investor behavior. Investors can be broadly classified into two groups (1) algorithmic-based high frequency traders or short-term players and (2) fundamentals-based low frequency traders or long-term players. Long-term players should be concerned with Corporate Risk or asset growth, i.e., if assets are growing the share price will increase. See Fig 2. Short-term players are primarily concerned with Expectation Risk.

Fig. 2: Depicts How Total Assets Determines Revenue, (Source: Ben Solomon, 2019)

Note: All data, as of January 2019, was obtained from Yahoo Finance. The symbols for each company analyzed in this research are:

1. Semicon: Texas Instruments TXN, Micron Technology MU, Intel INTC, Vishay VSH, Qualcomm QCOM, AMD AMD, Broadcom AVGO, Cypress Semiconductor Corporation CY, Apple AAPL,
2. Banks: Chase JPM, Bank of America BAC, Wells Fargo WFC, Citigroup C, US Bancorp USB-PO, Goldman Sachs GS, Morgan Stanley MS,
3. Retail: Kroger KR, Walmart WMT, Target TGT, SpartanNash SPTN, Amazon AMZN, Brown & Forman BF-B, Costco COST,
4. Apparel: The Gap GPS, American Eagle Outfitters AEO, Under Armour UAA, Capri Holdings Limited CPRI, Ralph Lauren RL, Abercrombie & Fitch ANF, Urban Outfitters URBN, The Children's Place PLCE.

Revenue Is Derived From Total Assets

Searching for a method to determine upside risk I plotted Revenue against Total Assets for four very different industries, Semicon, Banks, Retail and Apparel. Fig. 2, confirms that these are very different. I was shocked by the very strong regression relationships (look at the R²). Usually, R² above 90% is considered very good. In Social Sciences the joke is, the world is correlated (R²) at 30%. See, Appendix: A Word of Caution, for further information on regression analysis. Either, I am completely off the mark or this was a breakthrough and I had to research further.

The most important implication of Fig. 2 is that a company's ability, assuming well run, is determined by the total (current and long term) assets it has. That is, for every $B (billion) of assets, the Semicon industry generates $0.7166B in revenue, the Banks $0.078B, Retail $2.0109B and Apparel $0.7902B.

This makes logical business sense as, if you only have enough assets to produce 10 units of a product, you can only sell 10. Similarly, if you only have enough assets to produce 1,000,000 units you can sell 1,000,000 unless the market is saturated. This, market saturation, is the case with the top 100 brands like Proctor & Gamble (PG) and Unilever (UL). See "How Healthy Foods Are Nourishing Growth in the CPG Industry," "Top 100 CPG Brands Mostly Lost Sales And Share In Past Year," and "How CPG Bands And Retailers Can Remain Competitive Amid Seismic Change."

Liabilities and Equity Are Good Indicators of Revenue

So, what happened to the Liability/Equity side of the Balance Sheet? Doesn't the market care?

Fig. 3: Total Liabilities Determines Revenue, (Source: Ben Solomon, 2019)

Fig. 4: Equity Determines Revenue, (Source: Ben Solomon, 2019)

Figures 3 & 4 show similar relationships with Total Liabilities and Equity. This would be expected as Total Assets is the sum of Total Liabilities and Equity, i.e. they are strongly correlated but there is more noise in the Total Liabilities and Equity relationships. This is logically correct as revenue generation is not dependent upon how the assets are funded, whether cheaply or expensively.

Cost of funding is important for cash flow considerations. One could have 100% equity (as in sole proprietorships) with no debt or on the other extreme, 100% debt with no equity (not likely but theoretically possible). Most companies are somewhere in between. This needed further investigation.

Bear in mind that Liabilities and Equity don't generate revenue, only assets do, that is why I focus on Total Assets, to be as close as possible to source of Revenues.

Revenue Drives Market Capitalization

The next step was to determine whether there is a relationship between Market Capitalization and Revenue. I used the 3-month mid-range i.e. the difference between the highest and the lowest share price in the last 3 months (Yahoo Finance). This smooths out the data with a minimum algorithmic intervention, and at we are not concerned with the shape of the distribution of prices. It also matches the quarterly reporting. Fig. 5 shows the data graphically.

There is a very good correlation with Semicon and Banks, however, it is poor with Retail and Apparel. As you can see Amazon is way off in "left field" and throws the graph off. My inference was that Revenue and therefore, Total Assets (the revenue generator), was a very good predictor of Market Cap, but more digging was required. It appears that the market is not so much concerned with the cost of your liabilities (interest payments) or equity (dividend payments) as the market is only concerned with the bottom line totals, Total Assets and Total Liabilities & Equity.

Fig. 5: Revenue Determines Market Cap (Source: Ben Solomon, 2019)

The Engines of Growth

Changing our perspectives, from that of a company as a legal entity of ownership to that of a legal representation of its Economic Engine, valued by the market as Market Cap. To support this perspective change, the existence of markets prove that markets are not concerned with legal ownership as shareholder are always changing.

To fund these assets the company has a second engine, the Funding Engine. Much of the financial statement ratio analyses is conducted to determine how well this Funding Engine is running. Apparently. these analyses have nothing to do with the pricing of risk, risk of the share price.

As was observed with Old GM, its share price dived to penny stock status in less than a year. If my memory is correct Old GM lost $40+ in 8 months. Not because these financial ratios were not working but because Old GM's Economic Engine was not working. Its Economic Engine could no longer support the Funding Engine. Its Funding Engine was working fine, even with negative equity.

Is there an approach to determining this Economic Engine? Yes, after extensively pouring over and testing the data, I settled on a model that worked the best. First, there is no need to restate the relationships between assets and liabilities or equity as that is very well done with current financial statements analyses. I have no intention of replacing financial statements. The Economic Engine is represented by the Economic Statement and the Funding Engine is represented by the Funding Statement.

Funding Statements

Tables 1, 2, 3 & 4 depict the Funding Statements of the Semiconductor, Bank, Retail and Apparel companies. I have constructed the Funding Statement based on my experience as a Senior Credit Analyst in Commercial & Industrial lending. You can add more, and bankers could add covenants, but don't over burden yourself. As you can see it is a version of the financial statement ratio analysis, with one very important exception. All numbers are normalized by their respective Total Assets and thus, comparable across companies and industries.

Table 1: Semiconductor Funding Statement (Source: Ben Solomon, 2019)

Table 2: Banking Funding Statement (Source: Ben Solomon, 2019)

Table 3: Retail Funding Statement (Source: Ben Solomon, 2019)

Table 4: Apparel Funding Statement (Source: Ben Solomon, 2019)

I use the term "premium" as in insurance premium, to denote what a shareholder pays to keep both the Economic and Funding Engines running.

This Funding Premium, defined as (Operating Profit - Net Income) / Total Assets, includes debt interest payments and debt principal payments (unlike EBITDA) to lenders and dividend payments to shareholders. It is After Principal, Interest, Dividends, Before Taxes. For banks that use EBITDA the debt service coverage ratio is usually greater than or equal to 1.1 as EBITDA does not account for principal payments. This Funding Premium is what you pay to keep the Funding Engine running. For this analysis, I estimated 5% (20-year term) of principal, for principal payments. With linear regression this statistic introduces a linear shift (not noise) into the data which the regression modeling can handle.

The Funding Premium is equivalent to the Debt Service Coverage Ratio. The average Funding Premiums are 11.34% (Semicon), 1.59% (Banks), 10.77% (Retail) and 7.29% (Apparel). That is, comparatively Banks have very low leverage payments. Micron, Intel, Vishay, AMD and Apple have below average Funding Premiums, and therefore less dependent upon external sources of funding. Qualcomm is not doing well at 34.01%. In theory Goldman Sachs and Morgan Stanley could improve on their dependence on external sources of funding. Walmart, Amazon and Brown-Foreman are very dependent on their external sources of funding. American Eagle Outfitters and The Children's Place have exceptionally high Funding Premiums. For TI, Broadcom, Walmart, Brown-Forman, American Eagle Outfitters and The Children's Place this is due to very high dividend payments which is acceptable in the short term.

Economic Statements

As with the Funding Statements, all numbers in the Economic Statements are normalized by their respective Total Assets. Economic Premium is defined as the ratio of Revenue to Total Assets, or how well a company can convert assets into economically valuable revenue.

Table 5: Semiconductor Economic Statement (Source: Ben Solomon, 2019)

Table 6: Banking Economic Statement (Source: Ben Solomon, 2019)

Table 7: Retail Economic Statement (Source: Ben Solomon, 2019)

Table 8: Apparel Economic Statement (Source: Ben Solomon, 2019)

The Economic Premium for Banks is only 4.09% compared to 81.21%, 230.95% and 157.18% for Semicon, Retail and Apparel, respectively, and shows clearly how different these industries are. Intel (59.77%) and Broadcom (41.59%) have very low Economic Premiums. Qualcomm's Equity/Total Assets is only 2.84% and well below the industry's 43.40%. That is Qualcomm is having difficulty building equity. The reason is that its Funding Premium at 34.01% is excessively high compared to the industry's 11.34%. Its external sources of funds are negating its ability to build equity. I would recommend that Qualcomm is a sell. Ralph Lauren's (109.88%) and Hanesbrand (99.19%) Economic Premiums are very low and shows how the top and bottom brands are affected by their inputs.

I believe that Intel is a long-term hold. Here is why. TI earns above average Economic Premium or revenue/assets (91.77%) from its assets compared to Intel (59.77%), even though they are in the same industry segment with very similar Management Premiums.

With similar managements, this suggest that Intel has 32% (91.77% - 59.77%) revenue potential that it has not realized. See this very good article on Seeking Alpha: "Intel's 10nm Problems Have Implications Far Beyond What The Market Is Seeing." When Intel realizes its full capacity, its revenue potential will be $118B. That is, Intel is a long-term growth stock. Disclosure, I worked for Texas Instruments from 1983 to 1992 as a proprietary manufacturing software developer.

Yes, Intel may have some teething problems right now, but its chips are super fast. About 6 years ago, I was hunting for a better PC to buy. I have physics Excel models that take 8 hours to run. I stripped down one of these models and took it to Best Buy. Ran it on an Intel 4-core, and an AMD 6-core. I was shocked that the Intel's 4-core was about 50% faster than AMD's 6-core. I bought the Intel 6-core i7. My model run time dropped to 20 mins! You see, for people who need absolute speed there is no substitute for large monolithic chips, not even multiple smaller dies. Intel is a buy.

Process Premium is what shareholders are willing to tolerate to have a process that generates revenue, and is defined as the difference between Revenue and Gross Profit normalized by Total Assets. The Apparel business has a very large variation in Process Premium, and I believe this is largely due to brand and process structure, but more research is required. Obviously, the Retail businesses are process driven and have very high Process Premiums. If you do not have an exceptionally good process you cannot survive in this industry, but look at Amazon, its growth was achieved by having one of the lowest Process Premiums in the Retail industry. Brown-Forman is better than Amazons, but Amazon is a much more complex business. Brown-Forman is the benchmark here.

Caution needed here. Much is dependent on where companies allocate their costs, as can be seen in the banking industry. Goldman Sachs and Morgan Stanley reported substantial COGS, while the other banks reported $0. I have moved 5% of the Management Premium to COGS for these zero reporting banks. Why? If cost were really zero, customers would be showing up at bankers & tellers homes to deal with deposits and withdrawals. Don't you think so?

Management Premium, what shareholders are willing to tolerate a management team's decision-making ability, is defined as the difference between Gross Profit and Operating Profit normalized by Total Assets. The Retail and Apparel industries has very high Management Premiums at 39.06% and 53.42%, respectively, while Banks and Semicon are lower at 2.35% and 18.13%, respectively.

Forecasting Market Cap From Quarterly Statements

I then built and tested multivariate regression models (see Appendix: A Word of Caution) on this data to determine whether the Economic Engine was a useful concept i.e. could it determine market capitalization? I was very surprised. Getting R² greater than 99% for regression analyses is nearly unheard of unless that is how this phenomenon actually works.

Note, that the Adjusted R shows how much the model explains the data, the closer to 100% the better. Anything less than 75% needs a lot of reworking. Anything less than 30% is useless.

The Market Cap model takes the form (CD stands for Cash Dividends),

Market Cap (MC) = a.EP + b.PP + c.MP + d.FP + e.CD + i (1)

Table 9: Regression Results to Determine Market Cap (Source: Ben Solomon, 2019)

How Do You Interpret These Models?

The coefficient of EP, the Economic Premium, are all positive except for Apparel. This means that the market rewards for economic activity or how assets are converted into revenue except for Apparels where the market believes that Assets take away from economic activity. That is, the other Premiums are significantly more important for the Apparel industry.

The coefficient of PP, the Process Premiums, are all negative, except Apparel. This means that the market penalizes for any COGS activity and explains why full vertical integration is not necessarily, a good thing. Unless there are extenuating circumstances that necessitate vertical integration. From this perspective, AMD's breakup into AMD and Globalfoundaries was a good decision but as can be seen in Table 4, AMD's Process Premium is 91.28% and well able the other semicon companies. If AMD can lower this its share price will go up, but AMD has a long way to go.

Returning to Apparels. The signs are reversed for both Economic and Process Premiums. That is, the market rewards in-house processing at the expense of Assets, or the need for better Asset utilization and inventory management.

The coefficient of MP, the Management Premium, is positive for Semicon companies and negative for everyone else. The market is rewarding semicon companies for their management capability and penalizing everyone else. Apparel industry's Management is close to zero, suggesting that the market does not consider management interaction a valuable contribution to economic activity. This reflects a business structure that I have not investigated.

The coefficient of FP, the Funding Premium, is positive for Semicon companies - i.e., part of their share price is derived from Cash Dividend (CD) payments. That is, the market rewards Semicon company's dividend payments when it rewards for Management Premiums. However, for the other companies, the market penalizes for interest and dividend payments when Management Premium is negative. This makes sense as funding costs take away from the real economic value derived from assets. Again, Apparels are the exception as it is a strong positive. This implies that the market expects significant leverage from this industry. This can be seen from the strongly negative Cash Dividends coefficient, and in general the market does not place any value to Cash Dividends payments as both Banks and Retail have zero coefficients.

Note, that even though Amazon, Apple and Walmart are enormous companies, they still have similar economic characteristics as their smaller counterparts. That scaling is not rewarded by the market.

IPO, M&A, Angels, and VCs Managers Take Note

Table 10: Actual versus Model Market Cap / Total Assets (Source: Ben Solomon, 2019)

So, we have a model that works very well. Beware, these coefficients will change with time, with the changing economy and as companies migrate between sub-industries. I have highlighted in bold red where the errors are substantial.

In Semicon, Apple deviates substantially with actual Market Cap / Total Asset ratio at 2.39 versus the model's 0.47. Table 5 shows that surprisingly the market has given Apple's Management Premium an 8.46% compared to the industry's 18.13%. In fact, Apple's Economic Statement is very similar to Micron's, a basic/commodity memory products company. Therefore, there are two possible interpretations (1) its market cap is derived from the attention it gets from the media & press (excellent marketing capabilities) and therefore, not sustainable in the long term or (2) it is structurally different from other Semicon companies. This makes sense as Apple is not a pure semicon play.

In Banks, Chase (-9.31%) and BofA (+7.68%) have opposite error signs. Since major banks are practically identical Economic Engines, share price movement in the opposite directions are to be expect. However, this shows that you cannot diversify with bank stocks as they are identical. One bank stock is all you need to hold in your portfolio if diversification is your primary goal.

Retail has similar errors to Banks as Krogers (-29.59%) Walmart (+34.36%) and SpartanNash (-42.95%). Until I look into this further, my guess is that these errors are due to investor Expectation Risk.

Finally, in Apparel, Hanesbrands as an error of 318.03%. The Economic Statement, Table 8, shows that Hanesbrand has the lowest Management Premium (25.48%) compared to the industry (53.42%), suggesting that the market has the lowest confidence in Hanesbrand's management team.

IPO, M&A, Angel, and VC managers take note. It is now possible to independently estimate whether an IPO or an M&A is priced too high or too low.

So, What's the Market Risk?

Sure, there are lots of complicated mathematical formulations for risk. The key is to get the underlying concept right. The analysis so far shows that the two main components of market risk are Expectation Risk and Corporate Risk. In this article I will not address Corporate Risk.

Fig. 6: Share Price Determines Share Price Spread (Source: Ben Solomon, 2019)

I used the share price spread as difference between the 3-month high and low as a measure of the Expectation Risk. The share price used is the median of this 3-month high and low. Note I took out Apple and Amazon as these were extreme outliers.

Table 11 shows the regression results for each specific industry. The Expectation Risk modeled is,

Expectation Risk = Share Price Spread = a.Share_Price + i (2)

Table 11: Share Price Spread Regression Results (Source: Ben Solomon, 2019)

Expectation Risk as measured by share price spread - the variability in share prices - is purely a function of the share price for that specific industry. The greater the share price, the greater this spread. Table 11 shows that the Semicon industry has the least and Apparel the most Expectation Risk.

Levi Strauss Valuation

Let's consider what this Economic & Funding Statement approach has to say about the upcoming Levi Strauss IPO. Table 12 & 13 presents Levi Strauss Economic & Funding Statements, respectively, with the Apparel Industry results for ease of comparison.

Table 12: Levi Strauss Funding Statement (Source: Ben Solomon, 2019)

Table 13: Levi Strauss Economic Statement (Source: Ben Solomon, 2019)

Table 12 shows that Levi Strauss is a very well managed company. It is well immunized (NCA not funded by CL) and its Funding Premium is on the low side below industry average. Table 13 shows that its Net Economic Premium (EP-PP-MP-FP) is on the high side. It has one significant drawback, its Debt Coverage is weak.

Using the financials filed with the SEC and the Apparel parameters of Tables 9 & 11, the model results are presented in Tables 14 & 15.

Table 14: Model Market Cap & Share Price Results (Source: Ben Solomon, 2019)

Table 15: IPO Underwriter's Offer & Market's Expectation (Source: SEC & Wall St Journal, 2019)

Table 15 shows that the underwriter has proposed an IPO offer price of $16.00 per share to raise $674,666,672. However, Wall St reported (03/20/19) that the IPO offer will be $17/share leading to a valuation of $6.6B. Note that it is the job of the underwriters to get the maximum price that prospective shareholders will pay while at the same time ensuring all the outstanding shares are sold.

Table 14 shows that $17/share is not sustainable as the longer-term median share price of $9.96 with a low of $9.82 and a high of $10.11. Levi Strauss is definitely a short, given that it listed at $22.share.

Investor Expectation Trading Band?

OK, now that I have reviewed Levi Strauss, let's take another look at the 32 companies presented in this analysis. A trading band is the range between low and high, that a share price will fluctuate between. Table 16 shows the trading bands for the 32 companies.

For each of the four industry sectors, Table 16 presents (1) the median share price for a three-months period between October 2018 and January 2019, (2) the model range derived from Table 11, (3) model high, (4) model low, and (5) the share price as of 03/20/2019. We are interested in the bottom three rows of each industry sector.

Per the legend at the bottom of Table 16, the share prices that are shaded orange are within 10% of the model high. Those shaded green are within 10% of the low. Purple shade indicates that the share price has exceeded the model high, while blue of those that have gone below the model low. Unshaded share prices are within the model range by more than 10% on either side i.e. the 80% bandwidth.

Table 16 shows that 7 out of 32 times (approximately 20%) the share prices were outside the model range. These occurrences are about even for high and low given a relatively small sample size. That is, the model is not biased in either direction.

Being valid about 80% of the time shows that the bandwidth is reasonably broad and useful. For example, a very large bandwidth, say, between zero and infinity would not be useful, as it would not provide signals to act, either buy, sell, short or get ready to take action. This model is good and validates this approach to Expectation Risk.

Table 16: Trading Bands - Model Versus Actual (Source: Ben Solomon, 2019)

What to Take Away?

From the work I had done, earnings are noisy estimates of market capitalization. This is clearly evident from the Levi Strauss IPO at $17 by in part estimating earnings multiples versus a purely analytical forecasted range of $9.82 to $10.11. I believe that the IPO pricing is substantially overestimated. The Economic Statement approach provides a better and more reliable estimate of market cap and market risk. It is the approach to converting accounting data into economic data.

Appendix: A Word of Caution

For a great introduction into linear regression and multivariate linear regression lookup,

1. What is Simple Linear Regression?
2. An Introduction to Linear Regression Analysis (video)
3. Conducting a Multiple Regression using Microsoft Excel Data Analysis Tools (video)

A word of caution with statistical analyses,

1. Conceptual Model: It is very important to start with a logical model. In this article the logical model is equation (1). It took a while to get there as I had to develop a set of statements, Economic & Funding Statements, that would provide consistent data that was grounded in a company's Financial Statements, to build this logical model. Even then there were times I have had to exclude a data point or so.
2. Correlations: The independent variables may be correlated i.e. they are not exactly independent as they are related to each other to some degree. This will mess with your results as they show up as having low independent variable correlations. You need to use judgement to decide whether to keep or reject these independent variables. This is where your logical model is invaluable. Usually in economics and finance this is not a big problem.
3. Heteroscedasticity: Regression techniques are based on the assumption that the independent variable standard errors are Normally distributed and not heteroscedastic. Heteroscedasticity means that there is a relationship between the dependent variable and the standard errors of one or more independent variables i.e. the standard errors increase or decrease with the dependent variable. My recent paper "Interpreting the Universe's Expansion Redshift Data with Respect to Energy" explains how badly physics modeling can go wrong due to heteroscedasticity.
4. Statistic Background: To do this type of analysis requires at least a Master's level in statistics and many years using that skill. I have a 2-year Master's degree in Operations Research from the University of Lancaster, a British ivy league university, of which 50% is statistics. Unfortunately, most job posting for Data Analyst, require Python, C or some similar computer language but no mention of statistics. In my opinion these are Data Reporting jobs. To analyze data, you have to know the shape of the data and how it can or cannot be combined or transformed. Some years ago, I read an in-house report on the housing market. It looked very polished and sophisticated - to the untrained eye. To the trained eye it was garbage. I took it apart in minutes and lost my job. So be careful of the reports you read or write.
5. Finance Background: I believe that to do these types of market/corporate analyses requires a finance background, too, at least at a Master's level, as you will know the rules and more importantly know when to break them. (See here.)
6. Last word: When possible, leave statistics to the professionals.

Analyst’s Disclosure: I/we have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.

Comments

[Benjamin Solomon]

For all of you who disagreed with my work, I have been vindicated. My March/April 2019 valuation of LEVI at about $10 has been vindicated. The stock that was IPO'd at $22+ and hit $25+ soon after that, has seen a consistent and steady decline. It is now trading at $12+. If you did not heed my advice you would have lost about 50% of your principal investment.

[Benjamin Solomon]

Read @Mark Hibben "Qualcomm: The Party's Over" seekingalpha.com/...

Qualcomm' party is over. I bet most of you nay sayers lost money. Qualcomm's bad news is not over. Apple is most likely to renegotiate it's $4-$5 billion settlement.

[Robert Honeywill]

I do not understand these comments. At the date of this publication QCOM was ~$58 and has gone up to $66. Intel was ~$54 and has gone down to $45. Am I missing something?

[Livegood3]

You all didn't follow Benjamins instructions and sell...darn it, now how am I supposed to get QCOM in the 50s again!? LOL Long, to the moon!

[ryanwood22]

Well that aged badly (Your QCOM statement)

[mikefconnell]

Glad I didn't take that advice!

[Benjamin Solomon]

Another nail in the coffin.

The Wall Street Journal reported today, "A Surprising Connection Between the Bull Market and Stock Buybacks" (www.wsj.com/...),

"This comes as a surprise, because the conventional wisdom is that economic growth is by far the most important factor in the outperformance of certain countries’ stock markets in recent decades. The researchers found little support for this belief."

[Robert Honeywill]

Hi Benjamin, Thanks for an intriguing article. I might wish to quote you and include your Figure 1 in a future article, with source reference of course. Would that be okay with you?

[Benjamin Solomon]

@Robert Honeywill Thanks. Definitely

[Not so much]

"Determining Market Capitalization With EXTREME Consistency And Accuracy"?
Short Levi's - it's market cap is $9.82-$10.11
Perhaps you should pick a new profession?

[Not so much]

How did that short go on Levi's lol?

[Benjamin Solomon]

I read the comments – most are atrocious. What has happened to America? Newton’s 3rd Law of Motion does not apply – putting someone down is not going to raise you up. But the vindication is that with this third SA article I went from having 20+ actively investing followers to 111. Most contributors have less than 10. Wow! Good people don’t disparage, they quietly get on with their lives.

This article is an important read for 2 reasons. First, it shows that there is a lot of preparation before getting to data analysis. This can be split into 3 parts, (i) data preparation, (ii) data reporting and (iii) data analysis. Unfortunately, most people mistake data reporting for data analysis.

Mathematics is a very precise tool but beware of what the late Prof. Morris Klein said in his book “Mathematics: The Loss of Certainty” (www.amazon.com/...), that mathematics has become so sophisticated that it can be used to prove anything!

The fault in a precise mathematical model is that it sits on a wobbly (for the want of a fun term) conceptual model, and thus give inaccurate result with a great deal of precision. That was one of the points of my article that the Dividend Discount Model is wobbly. We need to search for something better. The E&F Statements I had proposed are just the first steps – using mathematics to consider what to investigate (arbitrage or fundamentals) and eliminate what not to (time waste).

Here is an exercise for you all. Try documenting your investment strategy in a rigorous manner. State what you know and what you don’t. How it affects your decisions. Is what you are doing sufficient? When do you use qualitative and when quantitative?

If you cannot do that then you are proof that luck exists.

[darturtle]

Pure financial analysts could not always accurately predict a tech company because once its underline technical strength is gone, the whole business crumble. Financial asset is a tool for it to fight but unless it acts correctly (not merely act), it will still sink.

INTC's key underline strength is gone --- its semi manufacturing leadership has gone last year. INTC cannot put more transistors into unit area as TSM and KR:005930. Worse, TSM works for many INTC's competitors. Giving proliferation of Android and ARM plus MSFT's support Windows 10 on ARM, INTC's another powerful weapon X86 ISA (share only with AMD) is not as powerful as before.

INTC needs to spend cash, lots of them to catch up with TSM by not worrying a few quarters of red inks which would piss off Wall Street. INTC also needs to do more .... Stop here, as we talk on stock speculation, let's not wish or expect but watch price-volume chart.

[Technosemi]

Wow. Just to go through this very rambling article would be a major job. Don't know what to think, but from the comments and my own view on, for example, QCOM, it seems there are many questions as to whether the way the data is presented and concluded upon is valid or not.

[mailmail]

A thesis based on 2000 years of daily real and simulated stock prices from 15 exchanges and indices around the world sounds impressive.

Out of genuine curiosity, through your research, what is the earliest daily stock price you were able to locate? What was the date on which this stock price occurred and on which exchange did it occur? How much of your information is simulated and how much is real?

Thank you for the article.

[Contrarianstrategies1]

"The bottom line is that, Intel, which used to be about 2 years ahead of competition in process technology, for the first time, is behind TSMC by at least a year if Intel can keep its 2019 HVM commitment. GF seems to have reached parity with Intel". These notes are from the seeking Alpha article from August 2018 that you brought up:"Intel's 10nm Problems Have Implications Far Beyond What The Market Is Seeing." Intel today is also well behind in the 5G race. How do you come up with the idea that intel has a revenue POTENTIAL of $118B ( 50% more then the current revenue), as a basis for buying and holding on to intel shares for the long haul. Sorry but your analysis comes through as a sad April Fools joke!.

[jst]

Everything you wrote is regurgitated nonsense. And to top it off, you posted this nonsense, as it has nothing to do with what the author researched and wrote about. In time Intel will surpass $118B in revenue. C.L.O.W.N.

[brainleft]

Qcom assets are not machines but rather the how to manual on 5g. I kept reading though since you possess obvious intelligence. statisticians can get lost in their oscillations, especially if they think they can guess future stock behavior which depends on what the shareholders and potential/or lack thereof shareholders decide to do. fear and greed

[DGrainger]

Too long.

Are you going on a book tour like M. O.?

[ron3637]

This is all above my pay grade.

[Gordon Shumway]

TLDR: trust me I’m a genius but the entire scientific community is too dumb to understand my revelations. Also dogs speak to me and I do my research on 4chan.

[Livegood3]

Exactly, April Fools Day guys!

[Nescience]

Happy April Fool's Day!

[zybexXL]

Quote:
"Yes, Intel may have some teething problems right now, but its chips are super fast. About 6 years ago, I was hunting for a better PC to buy. I have physics Excel models that take 8 hours to run. I stripped down one of these models and took it to Best Buy. Ran it on an Intel 4-core, and an AMD 6-core. I was shocked that the Intel's 4-core was about 50% faster than AMD's 6-core. I bought the Intel 6-core i7. My model run time dropped to 20 mins! You see, for people who need absolute speed there is no substitute for large monolithic chips, not even multiple smaller dies. Intel is a buy."
/quote

J.F.C. ... you call yourself an analyst?

[SVexpert]

Hahahha wtf is what i tought... what does a 6 year old pc chip from intel has to do with Qcoms 5g chip?!

[zybexXL]

Actually reading it now, it's an obvious April 1st troll article. Well done :)

[zybexXL]

... and following some of the links the author gives to his own literature/website... I do believe this guy seriously believes his claims (which include antigravity research, impossible physics, and LOTS of bogus credentials). Amazing self-promotion crap.