It’s been a long time since I’ve written anything but given the recent market upheaval (ostensibly related to the tapering and possible end of QE), I thought it might be a good time to dip my toe back into the water. Specifically I want to look at gold (GLD) and the idea that QE (or any of its various iterations) is inherently inflationary or hyper-inflationary due to the increase in the monetary base. This misconception has, in my opinion, led to gold being overvalued by approximately 30%.
I’m sure we all remember Grade 9 economics and the explanation of how fractional reserve banking works. The idea is pretty simple: a bank receives a $100 deposit from Client A and with a reserve ratio of 10% is allowed to lend 90% of it to Client B. Client B then deposits his $90 into another bank account and the bank is free to lend another $81 to Client C. In the end we end up with $900 held in client accounts and $100 held in reserve by the bank. So the obvious inference here is that the more money that the Federal Reserve prints (or as Mr. Bernanke likes to put it, “digitally creates”) the greater the supply of money in the banking system. In this particular case, 9 times greater (incidentally 10% is the required reserve ratio for the United States and as such this would be the expected multiplier for the US).
Since the start of the financial crisis the Federal Reserve’s balance sheet, and the monetary base, has grown by a factor of 3. Ignoring the natural run off (i.e. expiration of their existing securities) the Fed is adding $85 billion to the monetary base every month. The natural assumption would be that this massive increase in the monetary base must lead to a massive increase in the amounts held at depository institutions, ultimately leading to more lending, more dollars in the economy, and a decrease in the value of the dollar relative to other items in the economy (i.e. inflation). This, however, has proven to not really be the case as US banks have chosen to hold almost the entire amount in the form of excess reserves (see table below, since 1984 for dramatic effect).
As the chart shows, the increase in excess reserves almost exactly match the increase in the monetary base. So much so, in fact, that you could argue that QE is having no real impact on the actual monetary base. The question, of course, is why?
The answer is: regulations. What they don’t teach you in Grade 9 economics is that regulatory arbitrage is one of the most important aspects of banking and as such there is almost always a way around a rule. In the case of required reserves, since 1990 a depositary institution has not been required to keep any reserves on time and savings deposits (i.e. accounts that aren’t chequing accounts) or accounts owned by other than individuals (i.e. corporations). So if you buy a GIC or CD or even have a regular savings account (which had less than 6 transactions per month) the bank can lend out 100% of your deposit with no reserve requirement. As of Q1 2013 chequing accounts across the US only held $875 billion (total reserve requirement of just $87.5 billion) while time and savings deposits held $7.2 trillion (total reserve requirement of $0.00). Therefore, the most important actor in the creation of money is not the Federal Reserve, but rather depositary institutions (i.e. JPM, WFC, BAC, etc).
Banks are in the business of earning what they call net interest. This means they want to take your deposits (whether chequing or time and savings deposits) and pay you an amount (currently 0% on chequing or maybe as high as 1.5% on GICs) and then turn around and loan that money to another party at a higher rate. The banks also have the option of investing the money to try and earn a higher rate of return than their cost of capital. Prior to the introduction of risk weight capital adequacy in the 1990s banks were limited to a total asset base (including deposits) of some multiple of their tier 1 capital (see chart below, chart expresses capital as a percentage of total assets). Risk weights, unfortunately, just presented another opportunity for regulatory arbitrage and slowly the old multiple (tier 1 capital to total assets) expanded (as outlined in this speech by Thomas Hoenig). Ultimately this was partially (or wholly depending on who you ask) responsible for the 2008-2009 financial crisis.
As a result of the financial crisis most of the risk weights have been increased and Basel III was introduced (although not yet implemented). The result of which is stricter capital requirements (although still using a risk weighted system) and the re-introduction of an overall leverage ratio (similar to the one shown above). While this system has not yet been fully implemented in the United States, it is being introduced in parts. Every aspect of Basel III is going to increase capital requirements for banks over current requirements for risky assets and as such will actually contract the real money supply (all else equal). For example, just to hold the now $1.8 trillion they now hold in excess reserves (which is just cash sitting at the Fed) banks will need to have a minimum of $54 billion in Tier 1 capital (in addition to their other lending activities). This is the leverage ratio minimum. On a risk weighted basis the cash held at the Fed would be a 0% risk weight, but because banks will be subject to two separate leverage ratios going forward (both by regulators and by the market) only the most stringent is relevant.
The trillions of securities the Fed purchased were (most likely) all held at those financial institutions, so there is no net change in their assets or liabilities. The important thing to note is that they literally have no choice but to hold the funds in excess reserves UNLESS they increase their tangible common equity. The only way they can do that is through raising capital or retained earnings.
So what does all this mean? QE has had little to no impact on the growth of the real money supply (M2), in that without QE it is likely M2 would have shrunk substantially but with QE M2 has continued to grow at more or less its previous rate. This is because QE has taken risky assets off bank balance sheets and put them onto the Fed’s. This has helped to stabilize asset prices in a period where everyone has been deleveraging, but hasn’t caused rampant growth in money supply due to the primary limitation of money supply: bank capital. In fact, absent long periods of above average returns at banks, it is unlikely that we see M2 grow any faster than historical rates.
The above chart is indexed to the start of the recession (give or take). If you assume that the prices for everything leading into the recession were “fair”, then after the period of correction for the shocks to the monetary system, you would expect that everything would increase more or less in line with each other. You’ll notice that the Producer’s Price Index for commodities, the overall Consumer Price Index, and the S&P 500 all have taken wildly different paths but ended up at the same level and are all just slightly higher than when the recession began (I could have added more data points, like oil, but they all end in the same area as CPI, PPI, and the S&P). The one notable standout is gold, which is 70% higher (approximately) than when the recession began. So if your investment in gold had anything to do with perceived risk of higher inflation, it would be wise to note that gold is approximately 30% overvalued on that measure (even after recent dramatic sell-offs) and higher inflation will only come when and if banks start earning (and retaining) at much higher levels. It would seem that an investment in financials (XLF) or the index (SPY) would be much safer than GLD at these levels, as any tapering in QE will be seen as negative for GLD but will be the result of stronger economic data (positive for SPY/XLF).