There is no simple relationship between interest rates and the stock market. In particular, a lower interest rate doesn’t necessarily lead to a higher stock market and a higher interest rate doesn’t necessarily lead to a lower stock market.

The conventional wisdom on this topic is based largely on what happened over the past few decades. Clearly, US equities generally fared well after interest rates embarked on a long-term downward trend in the early 1980s and generally fared poorly when interest rates were in a rising trend during the 10-14 year period prior to the early 1980s. Also, the inverse relationship (a lower interest rate leads to a higher stock market and a higher interest rate leads to a lower stock market) seemed to make sense and was incorporated into a popular stock market valuation tool called “The Fed Model”.

The Fed Model compares the earnings yield of the S&P500 Index (the reciprocal of the S&P500′s P/E ratio, expressed as a percentage) with the 10-year T-Note yield to determine if the stock market is over-valued or under-valued. The higher the S&P500 yield relative to the 10-year T-Note yield, the better the value supposedly offered by the stock market. An implication is that if the 10-year yield is very low, the S&P500 can have a very high P/E ratio and still not be over-valued. For example, according to the Fed Model the S&P500 is attractively valued today. This is because even though the current P/E ratio is almost as high as it ever gets (excluding the 1999-2000 bubble period), the current earnings yield is well above the current 10-year T-Note yield.

However, the simple relationship between interest rates and the stock market only makes sense at a superficial level. It doesn’t hold up under deeper analysis. The reason is that the current value of a company is the sum of all of that company’s future cash flows discounted at some rate, and in most cases it will not be appropriate to use today’s interest rate to discount cash flows that won’t happen until many years or even decades into the future.

When picking a rate at which to discount distant cash flows it would be more reasonable to use a long-term average interest rate than to use the current interest rate. Furthermore, there is no good reason why the change in the interest rate over the next 12 months should significantly affect the interest rate used to discount cash flows that are expected to occur 10-20 years into the future.

But if it is wrong to assume that the stock market should trend inversely to the interest rate over long periods, then why did this assumption prove to be correct over the bulk of the past 50 years?

The first part of the answer is that over the very long term the stock market swings from under-valued to over-valued and back again and that in the early 1980s a bond market under-valuation extreme happened to coincide with a stock market under-valuation extreme. The second part of the answer is that financial market history goes back much further than 50 years and the simple relationship on which the Fed Model is based is not apparent prior to 1970. In essence, the theory that a lower interest rate leads to a higher stock market and a higher interest rate leads to a lower stock market is an artifact of the past 50 years.

The above statement is supported by the following charts. The charts show the Dow Industrials Index and the 10-year T-Note yield from the beginning of 1925 through to the end of 1968.

The US stock market (as represented by the Dow Industrials in this case) was in a secular bearish trend from 1929 until 1942. Apart from an upward spike due to fear of government default in 1931, the 10-year yield was in a downward trend during this bearish stock market period. The stock market then embarked on a secular bullish trend that didn’t end until the late-1960s. The 10-year yield was in an upward trend during this bullish stock market period. That is, the long-term relationship between interest rates and the stock market during 1929-1968 was the opposite of what it was over the past 50 years.

A large divergence between two fundamentally-correlated market prices is important because such a divergence usually will be closed via a big move in one or both prices. However, divergences sometimes build for an inconveniently long time before they start to matter.

The gold/commodity ratio (the US$ gold price divided by the GSCI Spot Commodity Index) and the T-Bond are strongly correlated over the long term. They also tend to be well correlated over shorter timeframes, but significant short-term divergences sometimes occur. One such divergence has been developing since the beginning of this year, with the T-Bond making a sequence of higher highs while the gold/commodity ratio stays below its late-December high. Note that even the recent surge to a new 5-year high by the US$ gold price was not enough to push the gold/GNX ratio above its late-December high.

The current divergence and previous similar divergences (higher highs for the T-Bond in parallel with lower highs for the gold/commodity ratio) are illustrated by the following chart. The previous similar divergences led to large declines in the T-Bond price and I can think of no reason to expect that it will be different this time.

[This post is an excerpt from a recent TSI commentary]

Gold tends to perform well relative to commodities in general when inflation expectations are FALLING. The evidence is presented below in chart form.

The first of the following charts shows the Expected CPI, which in this case is determined by subtracting the yield on the 5-year TIPS (Treasury Inflation Protected Security) from the yield on the 5-year T-Note. In effect, the chart shows the average annual “inflation” rate that the market expects the US government to report over the next 5 years. We’ve labeled all of the important highs and lows on this chart.

The second of the following charts shows the gold/commodity ratio (the US$ gold price divided by the GSCI Spot Commodity Index). The labels on this chart correspond to the labels on the first chart. For example, point A on the first chart is the same time as point A on the second chart.

Notice that in every case over the 6-year period covered by the following charts, a high for inflation expectations (the Expected CPI) is associated with a low for the gold/commodity ratio and a low for inflation expectations is associated with a high for the gold/commodity ratio.

The relationship we tried to show above is more clearly demonstrated by the next chart. On this chart the commodity/gold ratio (as opposed to the gold/commodity ratio) is compared to the ProShares Inflation Expectations ETF (RINF). The correlation clearly is strong, with commodities consistently outperforming gold when inflation expectations are rising and underperforming gold when inflation expectations are falling.

An implication of the above charts is that if inflation expectations are close to an intermediate-term bottom then the financial world is close to the start of a 6-12 month period during which the industrial metals perform better than gold. Alternatively, a further decline in inflation expectations (increasing fear of deflation) would lead to additional relative strength in gold.

We realise that the above message is the opposite of what most people believe about gold, but a lot of what most people believe about gold is not accurate. Of particular relevance to this discussion, gold has never been a hedge against “price inflation”.

Gold tends to perform relatively well during periods when financial-system and/or economic confidence is on the decline. The declining confidence sometimes will go hand-in-hand with rapid “price inflation”, but it isn’t reasonable to expect gold to be a useful hedge against what generally is considered these days to be normal “inflation”. In fact, part of the reason for the strong INVERSE relationship between the gold/commodity ratio and inflation expectations is the general view that “inflation” of 2%-3% is beneficial.

Our view is that the next three months could be dicey, especially if there’s another sharp decline in the stock market. However, we think that by the end of this year inflation expectations will be significantly higher and industrial metals such as copper and platinum will be significantly more expensive relative to gold.

[This post is a slightly-modified excerpt from a TSI commentary published about two weeks ago.]

The way that most new money was created over the past 10 years was different to how it was created during earlier cycles. During earlier cycles almost all new money was loaned into existence by commercial banks, but in the final few months of 2008 the Fed stopped relying on the commercial banks and began its own money-creation program (QE).

The difference is important because most of the money created by commercial banks is injected into the ‘real economy’ (the first receivers of the new money are businesses and the general public), whereas all of the money created by the Fed is injected into the financial markets (the first receivers of the new money are bond traders). The Fed’s new money eventually will find its way to Main Street (as opposed to Wall Street), but the rate of monetary inflation experienced by the ‘real economy’ during the years following the Global Financial Crisis was a lot lower than suggested by the change in the US True Money Supply (TMS). Consequently, there may have been a lot less mal-investment during the current cycle than during the years leading up to the 2007-2009 crisis.

Don’t get us wrong — there has been a huge amount of ill-conceived and misdirected investment due to the Fed’s money-pumping and associated suppression of interest rates. Due to these bad investments, corporate balance sheets are now much weaker, on average, than otherwise would be the case. In particular, the corporate world collectively has gone heavily into debt and in a lot of cases the debt has not been used productively. For example, it has been used to buy back shares or fund high-priced acquisitions. This will have very negative consequences for the stock market within the next few years, but wasting money on share buy-backs and over-paying for assets does not cause the business cycle.

The ‘boom’ phase of the business cycle happens when artificially-low interest rates prompt investment, on an economy-wide scale, in new production facilities and construction projects that would not have seemed viable in the absence of the distorted interest-rate signal. The ‘bust’ phase of the business cycle kicks off when it starts to become apparent that, due to rising construction/production costs and/or less consumer demand than forecast, the aforementioned investments either cannot be completed or will generate a lot less cash than originally expected. Widespread liquidation ensues, and — as long as policy-makers don’t do too much to ‘help’ — resources eventually get reallocated in a way that meshes with sustainable consumer demand. The economy recovers.

The above is background information for the following charts. The first chart shows the year-over-year (YOY) rate of change in US True Money Supply (TMS), a.k.a. the US monetary inflation rate. The second chart shows the US monetary inflation rate without the Fed’s direct additions and deletions*.

Note that the second chart does not show what would have happened to the US monetary inflation rate in the absence of the Fed. Regardless of whether the Fed is creating new money or not, it exerts a strong influence on the commercial banks. What we have tried to do with the second chart is isolate the monetary inflation that causes the business cycle.

Prior to late-2008 the charts are very similar, but from late-2008 onwards there are some big divergences. The most obvious divergence was in 2009, when the rate of growth in TMS extended the rapid upward trend that began in 2008 while the rate of growth in “TMS minus Fed” collapsed to well below zero. Also worth mentioning is that the rate of growth in “TMS minus Fed” was in negative territory from August-2013 to June-2014, a period during which the rate of growth in TMS never dropped below 7%.

The swings in the “TMS minus Fed” growth rate explain some of the important swings in the US economy. For example, the rapid increase in “TMS minus Fed” during 2011-2012 almost certainly is linked to the mad rush to invest in the shale oil industry, and the 2013-2014 plunge in “TMS minus Fed” would be partly responsible for the collapse of the shale-oil investment boom during 2014-2015. Although it was focused on a single industry, this was a classic case of the mal-investment that results in a boom-bust cycle.

Over the past 18 months the TMS growth rate has extended its major downward trend, but the “TMS minus Fed” growth rate has rebounded. This rebound could delay the start of a recession.

*We assume that the amount of money added by the Fed equals the increase in the Fed’s holdings of securities minus the increase in Reverse Purchase Agreements.