Google X: Let's Start Investing in Proven Innovators

Business Week recently published an amazing article about Google’s advanced research and development arm: Google X.  Google X is the successor to the Los Alamos labs of our grandparent’s generation, the Xerox PARC’s of the 1990’s and the NASA laboratories before the budget cuts of the last decade.  Google X is working on some pretty incredible and transformative projects: wearable computers (Google Glass), the first self-driving car, advanced nuclear reactors and some next-generation wind turbines.  The laboratory is led by the best and brightest that Google can find, including their co-founder and former CEO Sergey Brin.  Google brings these bright minds together and provides them with the resources and inspiration needed to drive innovation, and in return Google retains the intellectual property rights to their creations.  In a period where America is starved for innovation and questions about economic leadership are abound, it seems that there would be few better places to invest our resources than in R&D efforts at places like Google X.  So, how much is Google spending each year to achieve some of their truly inspiring results?  According to their annual report, Google spent just $6.8 billion on total R&D in 2012.  For context, that compares to the $140.8 billion spent by the Federal government on research and development.  So, what if there was a way to provide Google X with similar financial resources as the Federal government? What if we were able to provide Google X with a $50.0 billion annual R&D budget?

David Einhorn, the managing partner of the hedge fund Greenlight Capital, recently gave a presentation supporting the idea that Apple should issue large amounts of preferred stock in order to return cash to shareholders.  Einhorn is famous on Wall Street for his out-of-the-box and often contrarian investment picks, which are usually supported by an excruciatingly rigorous analysis and PowerPoint presentation.  With Apple, Einhorn argues that there is an enormous demand for “safe assets” given the great level of uncertainty in the markets, and that preferred shares issued by an enormous and well-respected multination corporation such as Apple could be as appealing, if not more appealing, to investors than buying ever increasing amounts of sovereign debt.  If investors are willing to turn over their savings to the Treasury, which has a dubious track record of producing any material commodity of significant value, then surely they would be willing to lend their savings to the corporation that has produced the iPod, iTunes, PowerBook, iPad and iPhone (maybe iTV?!)?  Einhorn goes on to say that Apple could use the proceeds from issuing these “iPrefs” in order to fund a dividend to common shareholders or other financial engineering approaches to enhance shareholder value.  But what if Apple, or other companies, didn’t use that money for financial engineering, but rather used it for actual engineering?

What if Google issued $50 billion worth of iPrefs each year in order to fund its R&D efforts at Google X?  Each iPref would pay a quarterly non-cash dividend of one share of Google common stock into perpetuity.  The value of the iPref would be determined by the expected future value of Google’s common stock.  Therefore, if the research work at Google X produced products that were commercially successful, then the iPrefs would receive a return on investment from their common shares.  In the interim, iPref investors would receive a steady dividend payment in the form of Google shares, which investors could then sell for cash on the market.  For investors who wanted a fixed cash return, rather than stock, they could simply enter into futures contracts with a broker: the investor would agree to sell the shares they received from their iPref dividends each quarter for a set period of time in exchange for a fixed dollar amount.  Based on current share prices and assuming a 5.0% dividend yield, my estimation is that issuing $50 billion in iPrefs would dilute Google’s common shareholders by less than 0.9% each year.  In other words, if Google X’s research projects were able to increase Google’s corporate profits by more than 0.9% annually, then both common and iPref investors would receive a positive return on investment. 

Why would an investor prefer an iPref instead of either a normal Google share or a bond issued by Google?  For starters, iPrefs would have no contractual default risk, which provides investors with certainty that their investment will not get locked up in bankruptcy court if Google ever ran into financial distress.  Secondly, the promise of new common share dividends into perpetuity means that the iPref investor will gradually accrue a larger ownership stake in the company, which would allow the investor to exert influence on the management of the company in the case that it was under-performing.  If the investor only owned common stock, then they would have to buy additional shares with additional investment in order to increase their ownership and influence.  Finally, the promise of share dividends into perpetuity would provide greater stability in the price of iPrefs relative to the underlying common stock, which would make them a superior store of value.

Most importantly, investors would seek Google (or Apple, or Microsoft, or Exon Mobil, or Berkshire Hathaway, or General Electric, or IBM, or Pfizer) iPrefs because their value is tied to Google’s ability to create and deliver products and services that consumers value.  Using iPrefs to transfer enormous amounts of our financial resources to these enterprises would allow them to take on much larger problems in order to develop more comprehensive solutions over much longer time-frames.  How quickly could Google field a self-driving car if it spent a $2.5 billion a year to have an additional 10,000 employees working on the project worldwide (Google has a total of 20,000 R&D employees today)? 

For a very long time we have been using financial innovation to promote unproductive economic activities, particularly in the higher education space.  Surely it’s about time to start reversing that trend.

Sen. Warren's Bill and Why We Need More Math Majors in the Senate

Massachusetts Senator Elizabeth Warren released her first authored legislation today: The Bank on Student Loans Fairness Act.  Essentially, Senator Warren argues that the Department of Education should lower interest rates on student loans to 0.75% because that is the rate at which the Federal Reserve System lends to deposit holding institutions (referred to as “Big Banks”).  Without treading too deeply into the political malaise, I’d like to take a little time to address concerns I have with the logical framework that Warren is relying on.

The Discount Window is a mechanism through which the Federal Reserve makes secured overnight loans to banks in exchange for high quality collateral (in most every instance, loan assets).  The Fed charges a pre-set interest rate, called the “discount rate”, on these overnight loans, which is what Warren refers to when she references a 0.75% interest rate.  However, the Fed also institutes a discount margin on these overnight loans; that is, the amount the Fed will lend is capped by a loan-to-value ratio based upon the quality of the underlying collateral.  This pre-set loan-to-value ratio is called the collateral margin (or the “haircut”) and is readily available on the Fed’s website: Fed Collateral Margins.

The only Big Bank able to borrow at 0.75% from the Fed is the US Treasury.  This is because Treasuries have a very low collateral margin, whereas most every other asset has a meaningful margin that caps borrowing.  For example, US Treasuries have a collateral margin of 97%, which means that the Fed will lend me $97M in cash at 0.75% interest for every $100M in Treasuries that I bring to the discount window.  Alternatively, AAA rated corporate bonds have a discount margin of 95%, so the Fed would only lend me $95M for $100M in AAA corporate bonds.  The collateral margin for private student loans is 83% because the Fed perceives them as significantly more risky than either Treasuries or AAA corporate bonds (which, clearly, they are).  Given that Federal student loans default at a rate that is roughly 5x higher than private student loans, we would expect the collateral margin for Federal student loans to be significantly less than 83%: let’s generously say 70%.

The rate a bank charges borrowers is dependent upon the collateral margin because loans that have a lower collateral margin are more expensive to fund.  For example, an investor can take in $5M in equity and use that equity to buy $100M in AAA corporate bonds: $5M comes from the equity and $95M comes from the Fed discount window.  If the AAA bonds pay 2.5% interest, then the investor will collect $2.5M in interest each year and pay just $0.7M in interest on the loans from the Fed ($95M x 0.75%), which makes them an annual profit of $1.8M or a 35% return on equity.  That’s a great deal!  Unfortunately, great deals don’t last very long, so soon competitors will jump in and start buying up AAA bonds, which will drive down yields until they reach a stable return on equity.  Today, AAA corporate bonds are paying 2.0% interest, which (under the same conditions) would provide a return on equity of 25%. 

How much would the Department of Education be able to charge in interest if it were lending under the same terms as the Big Banks?  To figure that out, we simply reverse the calculation we did above.  The investor (the DOE) needs to make a 25% return on equity in order to match the returns of the AAA corporate bonds.  However, in this case the investor can only borrow up to 70% of the loan value from the Fed, which means it needs to bring $30M in equity in order to purchase $100M in student loans.  Thus, the investor would need to charge an interest rate of over 8.0% in order to match the returns of the AAA corporate debt.  The figure below demonstrates how this math works for several different types of loans (note that we are not factoring in certain other risks that would drive interest rates even higher).  So, if Warren were to get her wish, then student loan interest should actually increase rather than decrease under her proposed financing scheme.

Now, all this is not to say that Warren is wrong in her belief that student loan interest rates are too high or that the federal government is unable to rectify the situation.  The Department of Education does not need to borrow from the Fed in order to make loans: it can borrow directly from the Treasury.  Therefore, the Department of Education can charge any interest rate it wants: it could charge 5.0% or 0.0% or even negative 5.0%.  However, these lower interest rates would produce investor (taxpayer) losses that probably should be accounted for. A simpler solution to lowering student loan burdens might be to stop making them in the first place.