There has been a down move in nominal prices in the equity markets recently, panic selling, margin calls, and massive money-printing responses from Central Banks ranging from outright buying all junk bonds , to monetizing equities by some non-us central banks (CBs)  .
I have speculated this will lead to a painful lesson in the near future: learning the difference between real gains (gains adjusted from inflation), and nominal gains (the nominal price quoted by the exchanges). In this post, I want to do an analysis of inflation and its effect on the market, as well a claim made recently by Mike Maloney that we entered a 60 year bear market, after the 1929 stock market crash, adjusted for inflation (using CPI-U)
Analyzing this data now is more important due to the recent, extreme, worldwide monetary policies of central banks, with permission, or in most cases encouragements of their governments that will have long-lasting economic repercussions for generations to come.
The Claim of the 60 Year Bear Market:
Click to enlarge
The stock market crash of 1929 historically was similar to today’s market, however, over the next 2.5 years the market dropped significantly (~90%) until it finally bottomed in 1932. Many historians said this lead to the Great Depression. Why the Great Depression lasted so long is debated and the subject of many research papers. I personally take the side that could it ended sooner, and it was so “Great” because of inappropriate government intervention, instead of having unfettered free markets (also eerily similar to today). Today we are seeing history repeat the mistakes of the past, perhaps even in greater magnitude than in 1929.
Recently I heard Mike Maloney state that after the Crash of 1929 it took the Dow Jones Industrial Average (DJIA) until 1992, or ~60 years in inflation adjusted dollars (using CPI-U), to reach a breakout high of the 1929 crash . Is this true? To verify this claim I had to use datasets (various sources) and software (QuantSpider) to validate this shocking claim.
I have researched SPX spot data back to 1927 (using the Yahoo OHLC data SPX, 1927 is when the dataset begins) when I was evaluating the SPXEPS strategy listed here on this site.
Trying to compare current history to the past can be difficult. Often the further you look back, when you are trying to compare market data, a lot of major variables can change. Also the data might not be as “clean,” as some quants can attest to, and getting this data can be not only difficult, but unreliable. Here are just some broad reasons, why underlying assumptions, even on looking at nominal prices, might not be as easy as you think:
- General Accepted Accounting Principles (GAAP) have changed significantly since 1927
- In 1927 dollars were denominated using a gold standard (currency backed by gold) (although the Fed was established in 1913), today we use fiat currency (currency backed by debt, or a promise)
- Markets since around 1998 have been digitized and decimalized
- Over the past decade or two, a record amount of central bank interventions are happening, worldwide, using Zero Interest Rate Policies (ZIRP) and Negative Interest Rate Policies (NIRP) policies. This has never happened in human history, dating back to even 600 B.C. ancient Greece and competing city-states.
And on and on.
When previously evaluating the SPXEPS strategy before 1981 it also did not perform well. One of the conclusion from the SPXEPS failure before 1980 was: the market needed to nominally move up much higher. In 1950 (~$20.00) -> 1980 (~$100), the SPX spot market only moved up by a nominal value of 500%. But since 1980 (~$100) to present (~$3,000), it has moved up approximately 3,000% in nominal value, a significant difference between these time periods to say the least. There are reasons I think this happened, but that is not the primary discussion of this post.
Click to enlarge
Mike Maloney makes a claim that the market entered a 63 year bear market from the 1929 crash, using the DJIA. This means, adjusted for all the all the money printing or inflation, you did not break even in the market, if you bought and held the high of 1929 (DJIA: 1929-09-03; SPX: 1929-09-06, or 18.03), until 1992 (DJIA: 1992-10-09; SPX: 1985-11-21, or 18.11).
(The market did not reach the great depression low until ~3 years later, or 1932-07-08, or a value of 3.18, or a drop of 82.3%, in real, inflation adjusted CPU-U terms.)
I wanted to validate this claim by looking at the SPX data (Mike Mahoney analyzed at the DJIA or Dow). I want to use the SPX, because while the Dow is not a bad metric, but it is just 30 stocks, where the S&P is 500 stocks, and generally captures 95% of the equity market. In general, the SPX is a better measure of determining the equity market.
Click to enlarge
You can see from the chart, it is not entirely true, in the 1958, and you did break even (SPX: 1958-12-11, $53.35 spot, or 18.09 in real), in real terms. So at face value, this is only half-true, and the true break-even in real terms, took only to the late 1950s roughly 30 years later to break even, in real terms, in the SPX market.
Click to enlarge
However, Mike is also correct, as we did enter a vicious bear market in real terms, from the 1970s inflation, so in real terms the SPX did not reach a true breakout in real terms, until in mid 80s (he claims 1992).
Click to enlarge
Problems with the CPI-U, methodology and changes:
The problem with using inflation adjusted data is that in the 1980s there were significant changes to the Consumer Price Index (CPI-U). CPI-U is an economic number derived monthly from the Bureau of Labor Statistics (BLS).
In fact, is it somewhat hard to find the changes, as there have been many, that were made to the CPI back in the 1980s, but needless to say there has been some debate to it .
- The BLS new methodology uses quality of goods substitution.
- John Williams calculates the original methodology based on a basket of goods having quantities and qualities fixed
- It now uses a cost of living index (COLI), instead of a cost of goods index (COGI) 
- The CPI-U, post-1980 uses consumer behavior, something rather subjective and non-numeric, rather than actually looking at spot prices
- Simply, consumers would drift toward lower quality of goods over time, yet CPI-U calculated inflation would not increase, even though prices of all goods are increasing. The BLS CPI-U would record 0% inflation, even though quality of goods is declining. Over time consumers choose substitutions of higher quality goods for lower quality goods [15, Meat Example]. Everything today seems to be made of cheap (China) plastic?
- This also explains a few years ago why shrinkflation was a economic topic. People would purchase the same item from the store shelf, yet it would have less quantity of goods. Shrinkflation is INFLATION. Inflation can easily be defined as: Buying less of a good or service with the same amount of currency; deflation being the opposite . Shrinkflation is inflation, but under the current methodology by the BLS, this is not considered inflationary.
For instance, the CPI’s medical spending are only weighted into 1% of the total CPI calculation today, yet medical spending accounts for 1/3 of all United States consumer spending, a huge difference on real (medical spending) vs. reported inflation (CPI-U) for the average person  .
Peter Schiff also did a thorough analysis showing discrepancies in the CPI-U, pre and post 1980, discussed in this video here :
Why did this change in CPI methodology occur? Well, I believe one key reason is a Conflict of Interest (COI). It is no coincidence that since the 1980s the debt has risen substantially as a % of our GDP . This is due to in large part from older demographics, an increasing aging population, and its resulting extension of government mandated entitlement benefits, now at an unimaginable sum of $255 trillion dollars .
The CPI is an extremely important calculation. It is used to calculate:
- Real Gross Domestic Product (GDP)
- Real bond yields
- Standard Tax deduction rates
- Social Security payouts (Another point of COI to understate CPI-U)
- Projected future cash flow models (important for business capital investment decisions)
One of the sources often uses by skeptics of the CPU-U is ShadowStats Alternative CPI-U dataset (SGS-CPI-U). William of ShadowStats doesn’t do the extensive analysis that Schiff did, but from what I researched William's adds a fixed percentage the to the CPI-U's understated number, without doing some of the transparent statistical calculations that Peter Schiff showed above.
William’s SGS-CPI-U proposed methodology is not without its legitimate criticisms  . I suspect the real CPI number in somewhere between the government number and the SGS number. It is a shame the methodology was changed. For example, understating CPI allows for lower bond yields in the market, as real rates are perceived higher by investors, using the government’s inflation number (CPI-U). Therefore it is advantageous for debtor nations to understate CPI, in order to keep bond yields low, and to issue more debt. The United States is the largest debtor nation in the history of the world, so it has a very strong COI in keeping interest payments as low as possible, as even now a meager interest rate of 4% (or sadly, even lower) on the 10-year bond could bankrupt the United States government.
Results from ShadowStats Alternative CPI number, or SGS-CPI-U:
As discussed above, Mike Maloney’s claims are true. Indeed the market took from 1929 to 1985 (or 56 years) to show any real value gain or breakout using the BLS’s CPI-U dataset. Mike Maloney did not analyze the spot equity markets using SGS-CPI-U dataset, an even more conservative dataset for inflation.
Click to enlarge
Using the SGS-CPI-U you can draw out that the “true breakout” from the 1929 crash didn’t happen until the 2009 bottom occurred, or SPX: 2009-07-03, $976.29 spot, or 18.46 real value. This would imply that the ‘bear market’ in real terms lasted 80 years.
In my opinion, the ShadowStats number makes more logical sense to use as the inflation adjustment dataset. As for reasons described above, since 1980: debt, inflation, world-wide money supply, and central bank balance sheets (ZIRP and NIRP) have all increased significantly. This has happened much more in recent years, and even more relevant today in these troubled times .
Quantitative easing (QE), or money printing, is a euphemism for inflation. It makes more sense to use the SGS dataset, because inflation does NOT CREATE REAL VALUE. History is littered with civilizations that did “QE” and failed . QE does not work, and is reflected, in my opinion, more accurately in the SGS dataset. If it did create value, governments from ancient times would have done this years, decades, centuries, or since the dawn of civilization itself, and succeeded; yet, “QE” has had a 100% fail rate.
This does NOT mean that nominal values of (equity) markets CANNOT increase, thereby having equity markets preserve purchasing power. But it does not INCREASE purchasing power, as we can see from the long-term spot SPX market since 1927 to present.
A major conclusion drawn is when applying the SGS-CPI-U dataset to the spot SPX number: Immaterial value in spot SPX has been created since inception of the Federal Reserve (SPX: 1927 | ~$18 -> SPX: 2020 ~$20), in terms of real prices (or value). These data results confirm the thesis: inflation does not create real value.
Problems of looking at the spot market, even adjusted for inflation:
A problem in the 60-year bear market CPI-U adjusted analysis is overlooking a key secret of the stock market: re-invested dividends.
The main problem with using spot SPX (or Dow), even though it is a widespread benchmark is:
- The price is effected by dividend payments, but It does not adjusted for dividends 
- The shareholder receives cash, the spot SPX is adjusted downward, but this cash received is not accounted, or adjusted backwards in the calculation for the SPX, resulting in a depressed index/share price the further back in time you go. This assumes the investor pockets the cash received from the SPX dividends, and the cash practically vanishes, and does not reinvest the dividends back into the SPX.
- SPX is a theoretical number, it not directly “tradable,” but it is a conglomerate of many companies.
- Until ETFs like $SPY and $VOO, the $SPX itself was not directly tradable. Granted now it has become much easier for casual investors to invest long-term in the SPX, even then there are management fees in the ETFs. This is why the SPY chart apperas much more bullish than the SPX chart, is because it is backwards adjusted for dividends.
- On an individual stock basis, there is an adjustment, but at the SPX level, there is no backwards adjustment
Adjusting for reinvested dividends:
- Take OHLC data from Yahoo finance of S&P 500 ($SPX) back to 1927.
- Use monthly SPX dividend dataset from MULTPL
- Annual rate is given monthly, then divided by 100 (convert from a percent), then divided by 12 (months / year). Then multiply this number by the closing share price on that day to give the theoretical monthly dollar dividend, re-invested.
- Adjust the price of SPX backwards using QuantSpider (MYSQL and PHP, software developed by myself) using the adjustment method
- Assume 100% of dividends are invested, with zero tax rate, or tax consequences applied to them. (Throughout US history this taxable percentage has changed, but even today there are ways to pay zero or low taxes on this by claiming residence in tax preferred provinces like Singapore, or Puerto Rico.) This will overestimate the dividend reinvested method by assuming no taxes were paid from the dividends. You could assume a 20% tax on dividends if you want to make an even more accurate assessment. Also in the past, brokerage fees, and spreads were much higher (slippage).
- Then inflation adjust the data to compare using CPI-U, and SGS-CPI-U, as above
- Compare and analyze the data.
Again, using the SGS-CPI-U inflation with the adjusted SPX, we see the more accurate picture of the real value of the market. In graph the number is rounded to 2 decimals as a final calculation, but it can go up to 10 decimal places in the software.
Click to enlarge
- As of 1929-09-06, the SGS-CPI-U, dividend adjusted price is $0.61. When accounting for dividends (which have been historically much higher than today)
- The low was put in on 1932-06-01 at $0.13
- It only took the SPX to 1936-10-06 to reach the high of the 1929 crash, (CPI-U adjusted, because SGS doesn’t start till 1980-01-01), or in 7 years, far earlier than the 2009, or 1990s, or 1980s, or 1960s as stated before (depending on the dataset used and if dividend adjusted).
- We did not experience a “true” breakout until 1945-01-08, when I’m sure, victory in WW2 was assured
- Hitler committed suicide on 1945-04-30
- Germany surrendered on 1945-05-07
- Hiroshima happened on 1945-08-06
Click to enlarge
Click to enlarge
Other interesting facts of note:
- Volume, even inflation SGS-CPI-U adjusted, picked up substantially over the life of the market
- Volume data is not a part of the dataset until 1950, which it was 55K, today it is ~35M, or up a factor of 636 times (recently could be due to HFTs).
- The dataset ends at $25.72, and begins at $0.32, roughly a gain of 8038% SGS-CPI-U inflation adjusted, if dividends were reinvested at a 0% tax rate over that lifetime of the asset. In other words, the market (or dare I say, [world] economy) has grown by a factor of 80 times in real terms since 1927 to today (2020-04-20), using even the most conservative inflation measures available.
- There are long periods where bear markets exist, even when adjusting for dividends and inflation.
- 1973 -> 1984
- 2000 -> Present
- Since 2000 there has been heavy Fed intervention. This is once again, another argument for free markets, as the data shows, we have been in a bear market in equities since 2000 (the dow/gold ratio also confirms this).
- Nominal prices (unadjusted, raw spot SPX) differ greatly from the non-adjusted SGS-CPI-U number (practically flat over the entire life of the market). This means almost 100% of the value created from the market comes from dividends that are re-invested.
While Mike Maloney was correct, he did not account for a key market secret: Almost all real value is created in the markets is re-invested dividends, even when using a stricter inflation dataset.
This is the main takeaway from this article. Yes, we know the market has gone up nominally, but in real terms, almost all value is created from re-invested dividends. Inflation has created no value, and has transfered value from low to middle class people to rich people by at least double since 1989, and real wages have not increases since 1970 (since we left the gold standard). It has pushed up asset prices (inflation), and rich people are typically the ones with the staying power to hold assets, and re-invest those dividends.
Another key takeaway is since the real value (or inflation adjusted price) of the market is practically flat over the entire life of the market, there is almost zero benefit in timing the market. And while the real market value might remain depressed (as it has since 2000), over long stretches of time, re-investing dividends, does add real value; yet it is still advantageous to buy when prices are lower.
When using the CPI-U, rather than the SGS-CPI-U to factor in inflation, results differ significantly since 1980. If we assume the governments’ number is correct, we are in fact significantly higher in real terms, when the average person knows this is not true, as they are, for example, substituting higher quality goods to lower quality goods. I speculate the CPI-U number in the future will be highly debated and its validity challenged, as the Fed is currently doing QE infinity.
Debt through inflation has increased substantially, and what appears to have happened is debt is trying to replace value since the year 2000; this is another thesis that should be explored further. The SGS-CPI-U number reflects this idea more, as unproductive debt (especially debt created by inflation) does not create real value. To reverse this paradigm would require an extreme change in mindset and approach, which I am not observing at this time, but rather the opposite: QE infinity.
The charts were created from QuantSpider.com; I am still programming features of the software and it is still in development. It is not ready to for a full release at this time, but has a lot of features, and even more features are planned. Sign up for free now if you want, as I will be improving it over time. It is a powerful tool that will allow investors to validate strategies, as well as re-create and analysis the charts listed in this blog post. I hope you will be able to access it to validate my own work, and feel free to leave comments or criticisms below.