Recently, I have been thinking a lot about the US Fiscal Cliff, and the various issues surrounding it. As most SA readers are undoubtedly aware, one of the primary sticking points in the negotiations is the raising of the top marginal tax rate on high income earners. The Democrats, as always, support this tax increase, while the Republicans rather predictably oppose it. After considering this issue for some time, and analyzing as much data as I could find, I have some to two conclusions that I believe are salient and investable.
1) Rates on high income earners are essentially guaranteed to go up next year. All the Democrats have to do to achieve this is allow the Bush Tax Cuts to expire, and refuse to pass them again. This means that I believe the chance of "going over" the Fiscal Cliff approaches 100%.
2) While this will cause the markets to sell off in the short term (alongside declining corporate earnings and revenue), after they have bottomed out, I believe that there will be substantially greater reasons to be bullish about future earnings and revenue after the drop. This bottoming process could be severe, and may take several quarters or even years to finish before a true bull market begins again.
Conclusion #1 is a simple judgment call of mine, based on the policy objectives of the Democrats and Republicans and what I believe to be their relative positions of power. The Republicans are negotiating from a position of extreme weakness, and in fact will suffer more political damage no matter what they do. If they give in and agree to a raise in the top marginal tax bracket rate, they will be seen as having lost. If they hold out and no deal is reached, Republican intransigence will largely be seen as the cause, and they will be forced to pass a new tax bill that leaves the top marginal tax rate elevated. My opinion is that a deal will be reached fairly quickly in 2013, but we will go over the Fiscal Cliff briefly.
I will spend the remainder of this article, and a follow-up article, explaining the reasoning and data analysis that have led me to Conclusion #2. I will also outline the framework for a simple but effective economic model that can be used to predict future broad market direction.
We'll begin by going over historical US economic data. I've chosen the S&P 500 (SPY) as a proxy for all of the markets, and for economic productivity in general. I think having as much data as possible is helpful, so I've gone all the way back to 1960. The metric of choice that I've chosen to focus on within the S&P 500 is Annual Earnings. Earnings are ultimately what drive the valuation of equity assets, by such metrics as the Price to Earnings Ratio (P/E). As such, I believe a focus on them is appropriate. It is also important to remember that, to be sustainable, earnings should be supported by strong revenue. While margin expansion of the whole market can boost earnings for a short time, it cannot be sustained. So we will be examining S&P 500 earnings throughout this article and the next, keeping in consideration the idea that earnings are driven by revenue.
The chart above shows that there has been growth of S&P 500 earnings in dollar terms since 1960. We may note that, while it has not risen smoothly, the general trend is upwards, and earnings of the index have multiplied about 30X since 1960. This looks pretty good at first glance. However, I think it is important to keep in mind what is driving these earnings. Revenue has driven these earnings, over time. Well, then, what drives the increases in revenue? I believe that this can be broken into a number of factors.
1) Inflation. This is perhaps the most obvious of the factors I have identified in pushing up S&P 500 earnings over time. Those dollars are worth less every year, so the trend isn't quite as strong in real terms as in dollar terms. You can see above a chart of S&P 500 annual earnings since 1960 adjusted for inflation, by reporting all values in 2005 dollars. You can see that the uptrend is still present, but the increase in real terms is only about 5X, rather than 30X.
2) Population. This is something that many people may not consider when trying to predict future trends in the markets, but it is vitally important. More people is going to translate to more spending, which will translate to more corporate revenues and earnings. When people have babies, the markets must go up. The above chart adjusts not only for US dollar inflation, but also for increasing population over the years. A 1:1 linear relationship is assumed between population and S&P 500 corporate earnings. Since the S&P 500 is considered a global index, the world population figure for each year was used to make the adjustment, but the US population figure is not much different. We can now see that S&P 500 earnings are down to only about a 2X gain since 1960 when adjusted for inflation and population increase. What else is driving the earnings?
(click to enlarge)
3) Productivity. This is another thing that isn't always considered in the context of the stock market, but it has a big effect on earnings. Being able to produce the same good or service for half the input cost should lead to large earnings increases, ceteris paribus. In constructing the above chart, I've made a few assumptions about the effect of productivity gains on the markets. First, the data I used for actual productivity values were for US Industrial Agricultural total productivity. I believe that this is a good proxy for productivity gains as a whole. The second assumption I made was that increased productivity would impact earnings in a 1:1 fashion, being an effect that would presumably target net profit margins. The data when adjusted in this fashion show that we were in 2011 at the same place that we began in 1960 in terms of adjusted S&P 500 earnings for population, inflation, and productivity. However, the average of the last 5 years is actually considerably lower than the average of the years 1960-1964. Thus, we see that inflation, population increase, and productivity increase are having somewhat less than their expected effect on corporate earnings. If we have successfully accounted for all influences on corporate earnings, they should be completely flat at this point. This is not the case, and moreover we have one more big factor to consider.
4) Government Deficits and Debt Spending. The chart above gives a picture of what percentage of the aggregate US Economy, as measured by annual GDP, has been financed by debt spending. As with other data, the chart contains a number of assumptions which should be considered for reasonableness by the reader. The chart contains three kinds of debt increases - household, corporate, and sovereign. In the case of household debt, yoy debt increases were used to compute an additional yearly amount of debt spending. Corporate debt increases were modeled as a simple 1.15X household debt increase, since the amount of US corporate debt is about that much bigger than US household debt as of now. It should be considered that not all debt is equal. Debt which is used to buy productive assets that generate return superior to the cost of servicing the debt is good, while debt which does not do this is bad. It's difficult to say how much of each kind we have. Some debt, such as the mortgage debt taken on at the height of the bubble, is far greater than the true value of the underlying asset, even. So that is really bad debt. So is debt used to buy assets for which there is no longer much of a market. Current examples of this would be Solyndra, or Appalachian coal mines, or Oil Tankers. I've taken the bottom line approach in the model that debt spending should simply not count at all towards generating real earnings over time, and thus divided it out of my model. Thus, periods in which not much debt was taken on but spending (corporate revenues!) was driven from savings and income will show higher adjusted earnings, while periods in which spending was driven by taking on more debt will show lower adjusted earnings.
As we can see, removing the effect on earnings of debt spending does not change the historical S&P 500 earnings a great deal, but the downward adjusted trend is slightly accentuated. I think it is also very important to consider aggregate levels of current debt as compared to the past and consider what they bode for future spending. US household debt at present is close to 100% of US annual GDP.
Debt, whether sovereign, household, or corporate, cannot be increased forever without bad consequences. Added leverage carries interest and other service costs, increases volatility and uncertainty, and ultimately must be removed via a deleveraging process down to more sustainable levels, which imposes additional costs on growth. Our historically high levels of US household debt are terrifying on the surface, but merely worrying when gains in inflation, population, and productivity are taken into account. I believe that it is significant that adjusted US household debt is seen to rise since 1960, while adjusted S&P 500 earnings fall. Yet I think the falling of the adjusted earnings has an even more immediate cause, and one that will start to be addressed by the new US tax policy that I believe we will see in 2013.
In Part II of this article, I will lay out a case for what I believe to be the ultimate cause of declining adjusted S&P 500 earnings, and why a higher top marginal tax rate will be bullish for the markets.