Summary: Based on historical trends, three key aggregated stock market ratios – price/earnings, price/sales, and price as a percent of GDP – all show that publicly traded U.S. stock are overvalued by approximately 50%. This article explains the significance of these ratios, then, using a financial model developed specifically for this purpose, evaluates the impact of a major stock market correction on the funded status of California’s pension funds.
For all of California’s pension funds consolidated, the analysis finds that if the market corrects downwards by 50%, then recovers to grow at 5% per year, without benefit reductions, the required annual contribution will rise to 80% of pension eligible payroll. This will cost California’s taxpayers an additional $50 billion per year. If the market corrects by 50% then recovers to grow at 4% per year, without benefit reductions, the required annual contribution will rise to 113% of pension eligible payroll. This scenario will cost California’s taxpayers an additional $86 billion per year. The study evaluates the impact of benefit reductions, exploring cases where only new employees are affected, as well as cases where benefits earned for future work by existing employees are also affected.
When evaluating what level of benefit reductions can preserve pension systems without increasing employer contributions – which currently average about 33% of pension-eligible payroll – the study found the following:
In the case of a stock market negative correction of 50% followed by 5% annual growth, if the annual multiplier (the amount that is multiplied by years work times final salary to calculate an initial pension) is cut by 40% for new employees, the annual retiree cost-of-living adjustment is cut by 40% for new employees, the retirement eligibility age is raised by five years for new employees, AND the annual multiplier is cut by 20% for existing employees for future work, the current level of payments – approx. 33% – would be sufficient to bring the systems to 100% funded status by 2050.
In the case of a stock market negative correction of 50% followed by 4% annual growth, if the annual multiplier is cut by 40% for new employees AND existing employees for future work, and the annual retiree cost-of-living adjustment is cut by 40% for new employees AND existing employees, and the retirement eligibility age is raised by five years for new employees AND existing employees, the current level of payments – approx 33% – would be sufficient to bring the systems to 100% funded status by 2050.
The article concludes by recommending policymakers and stakeholders negotiate contingency plans to handle a major stock market correction that strikes an appropriate balance between costly tax increases and benefit reductions. The model used for this analysis can be downloaded here. Researchers at the California Policy Center may be available to assist those interested in using this model to analyze their particular pension system.
Pension systems rely on investment returns to ensure their ability to pay retired participants far more in retirement benefits than they, along with their employers, ever contributed into those funds during the times they were working. Public employee pension systems are particularly dependent on strong investment returns, because they are exempt from ERISA guidelines that mandate conservative, low-risk investments. Pension systems subject to ERISA are forced by its rules to make modest return-on-investment projections, and as a result they collect relatively more from their participants and their employers during their working years, and pay out relatively less in retirement benefits.
Because public employee pension systems make higher risk investments and depend on higher returns – in order to collect less from their participants and pay out more in retirement benefits – they are unusually vulnerable to sustained downturns in the stock markets. These pension funds typically have 70% or more of their assets invested in publicly traded stocks, and these investments are typically where they expect to earn the bulk of their high returns.
California’s public employee pension systems currently project between 7.0% and 7.5% per year on their investments. Some of the major funds have announced that over the next 20 years they intend to drop that projection to 6.5%. They adhere to this rather bullish long-term projection despite a stock market that in aggregate has returned almost nothing to its investors for nearly two years. Because public sector pension funds are so big, managing in aggregate over $4.0 trillion in assets nationwide, they cannot reasonably expect to beat the market. One of the premises of this article is that the historical performance of the stock market over the past 60 years, and especially over the past 30 years, is not an indication of similar returns from now on, and may in fact be counter-indicative.
This article will discuss the probability of a major market correction in part one. It will present a financial model to analyze the impact of this correction in part two, and it will present various options to policymakers in part three. It is important to emphasize from the outset, however, that there are no easy solutions to public pensions in the face of major downward correction in the market followed by a recovery that takes several years.
Last month we released a related article entitled “The Coming Public Pension Apocalypse, and What to Do About It.” That article is recommended reading for anyone interested in this topic. The format of this article parallels the earlier work but uses a different but complementary analysis in part one, and then relies on a comprehensive pension analysis model – prepared over the past two months and publicly introduced here – as a framework for discussion in parts two and three.
PART ONE: USING KEY RATIOS TO ESTIMATE HOW MUCH THE MARKETS ARE OVERVALUED
In the previous article, part one surveyed the reasons why pension fund rates of return will fall dramatically. The emphasis was on causes, which can be briefly summarized in two categories, demographics and debt.
Causes of unsustainable stock market appreciation: Demographics and Debt
The challenge demographics poses to pension funds is based on the rapidly changing ratio of retirees to workers. According to the Federal Administration on Aging, using Census Bureau data, in 1970, 9.9% of the US population was over the age of 65. By 2030 that percentage is projected to more than double, to 19.3% of the U.S. population. This means that twice as many people, as a percent of the total population, will be retired and living on their savings. Simply put, this means they will be sellers instead of buyers in the market, driving down prices. Pension funds, where you would not expect such a dramatic change in the ratio of workers to retirees since their demographics are based on growth of the government workforce, not birth-rates, have also recently become net sellers in the markets, because the benefit enhancements they lobbied for starting around 1999 are now translating into a retired population of who are people retiring earlier with larger pensions compared to older pensioners.
The challenge rising levels of debt poses to pension funds is based on three related trends – interest rates, total market debt as a percent of GDP, and stock market appreciation. What has happened since 1980 has been a long-term downward trend in interest rates, a long-term upward trend in the stock market, and a long term upward trend in total debt as a percent of GDP. These trends are connected: Lowering interest rates stimulates borrowing which increases consumer debt and creates an incentive (fully realized) for increasing government debt. Each time interest rates are lowered, two things happen to the stock market, (1) corporate profits rise because new rounds of borrowing stimulate consumer and government spending, which in turn makes their stock worth more, and (2) more people are buying stock since the return they can get on loans has just gone down – also driving up stock prices. The problem with this model however is that it is unsustainable. Ever since the last stock market low in February 2009, interest rates have remained at or near zero, and debt as a percent of GDP has plateaued at roughly 350%. Lowering interest rates can no longer stimulate economic activity or stock market appreciation.
Evidence of unsustainable stock market appreciation: Three Key Ratios
(1) Aggregate Price/Earnings Ratio
The first table illustrates the historical trend for the price-to-earnings ratio for the S&P 500. The historical average price earnings ratio for publicly traded stocks in the U.S., as can be seen from looking at the trend line, is about 16x. Currently it is about 25x. It should be noted that while earnings are considered too volatile to serve as the most reliable indicator of whether or not a stock is overvalued or undervalued, the table here, developed by economist Robert Shiller, presents a “cyclically adjusted price to earnings ratio.” Using this technique, a rolling ten-year average earnings is calculated, and that is what is being depicted on the chart.
As can be seen, the stock market crash of 1929 was preceded by an aggregate P/E of 30. The overvalued stock market before the correction of 2000, soaring on internet IPOs for companies that, in many cases had no earnings, achieved a P/E of nearly 45 before the fall. And the overvalued stock market before the correction of 2009 had a P/E of just over 25. These high P/Es are not sustainable. The S&P 500 closed on July 8th, 2016 at a value of 2,129. If the P/E of the S&P 500 were to revert to a sustainable historical average of 16, the S&P index would fall by 36% to 1,362. And while that’s a usable projection, in reality when corrections occur, the P/E ratio falls below the average for a time, i.e., a correction to 1,362 might be a best-case scenario for the S&P 500.
(2) Aggregate Price/Sales Ratio
If the price/earnings ratio, even smoothed to a ten year rolling average, is considered unreliable because earnings are volatile, then the lesser known price/sales ratio is a good way to take a second look. A price/sales ratio is relatively easy to grasp intuitively – it measures the annual revenue of the company divided by the value of the company. An aggregate price/sales ratio for the entire S&P 500 would simply be the total annual revenues of all 500 companies, divided by the total market value (expressed in their stock price x number of shares outstanding) of all 500 companies. The sustainable historical average price/sales ratio for the S&P 500 is 0.9x.
The next chart, going back to 1965, shows that aggregate price/sales ratios for the S&P 500 are at 50 year highs. As can be seen, at the 2000 peak the price/sales ratio was around 1.6x, and back in 2008 it moved just over 1.8x. Today the aggregate price/sales ratio for the S&P 500 is about 2.1x. If the P/S of the S&P 500 were to revert to a sustainable historical average of 0.9x, the S&P index would fall by 57% to 912.
(3) Market value of all publicly traded U.S. stocks/GDP
Another way, using ratios, to assess whether or not the U.S. stock market is overvalued is to compare the market value of all publicly traded U.S. stocks to the gross domestic product for the U.S. economy. The chart depicting the trends for this ratio since 1955 shows that it is just off a 60 year high. Before the 2000 correction the market cap/GDP ratio was 204%. Before the correction of 2009 it got up to 183%. And at the end of 2015 it was just under 200%. After both the 2000 and the 2009 correction the market cap/GDP ratio dove to around 100%. If public stocks in the U.S. were to revert to a market cap/GDP ratio of 100%, the S&P 500 would fall 50% to 1,064.
Market Value of all Publicly Traded U.S. Stocks / GDP
Last 60 years through 2015
The conclusion of this section, using the average of the three ratios just considered, is that if stock prices were to return to sustainable levels, the indexes would fall by 47%. This is consistent with the findings of the previous article, which demonstrated that debt formation can no longer be used to stimulate the economy and the markets because interest rates have been stuck at or near zero for the last seven years. The next section will examine the impact of a 47% correction on the pension funds.
PART TWO: THE IMPACT OF A MAJOR MARKET CORRECTION ON CALIFORNIA’S PENSION FUNDS
To evaluate the impact of a major market correction on California’s state/local pension systems, we have created a model to perform “Pension System Long-Range Financial Analysis.” As will be demonstrated, this model projects the annual cash flow for a pension system through 2075 under a variety of assumptions. It also shows the assets, liabilities (NPV) and funding status for a pension system by year through 2075. Readers are invited to download this model, evaluate its internal logic, view the assumptions underlying the baseline case, and experiment with various scenarios.
Download model for “Pension System Long-Range Financial Analysis.”
Baseline Case: No change to pension benefits, no reduction in return on investments
As a starting point for comparisons, the model projects the future cash flow and funding status for all of California’s state and local government worker pension systems combined under the following assumptions:
- Contributions are not increased.
- Benefits are not reduced.
- Annual returns on invested assets, with the exception of 2015 which we optimistically estimate yielded 2.0%, continue from 2016 through 2075 at 7.5% per year without exception.
Key cash flow and benefit assumptions input in this baseline case include the following:
- The average participant works 25 years, and there is a five year gap, on average, between when the average participant retires, and when they become eligible for pension benefits.
- The average benefit multiplier (amount that is multiplied by years worked times final salary to calculate the initial amount of a pension) for retirees through 1999 is 1.8%.
- The average benefit multiplier for post 1999 retirees is 2.5%, to reflect the retroactive benefit enhancements that began in that year.
- The average pension cost-of-living increase per year is 2.5%.
The model permits variations to the inputs that are beyond the scope of this report to describe in full, but retains sufficient simplicity that anyone downloading it with a reasonable understanding of spreadsheets and pension finance can examine them in detail. Comments and criticisms on the efficacy of this model are most welcome. To summarize a few of these features, the user can:
- Vary the life expectancy – we have assumed a life expectancy increase of 0.15% per year,
- Vary the number of workers according to the birth year of each participant – we have used Census Bureau data on California’s state/local workforce and state/local retirees to assume a roughly accurate total of 1.8 million pension “full-time equivalent” eligible workers and 1.8 million “full career” retirees, straight-lined at 60,000 participants per year of birth.
- Vary the amount that the beginning salary of a worker increases each year – we have assumed 2.75% per year which translates into an average starting salary of $15,000 in 1960 becomes an average starting salary of $66K in 2015.
- Change the amount of annual salary increase (promotion, “step increases,” and COLAs) – we have assumed 3.0% per year which translates into an average salary of $71K in 2015 for a worker hired in 1990. US Census estimates place the average salary for a state/local government worker in California at $72K. Our average pension eligible salary, using the above-noted assumptions, is $69K, which assumes an average overtime payment (not pension eligible) of 5%.
By inputting these assumptions and others into the model, it calculates the following results for California’s consolidated state/local pension systems at the end of 2015: A funded ratio of 65%, total assets of $756 billion, total liabilities (NPV) of $1.16 trillion, incoming contributions of $34 billion, and contributions as a percent of payroll of 32%. These results are almost exactly consistent with the consolidated estimates as provided by the U.S. Census Bureau for 2014 and 2015. Those references are footnoted on the “data” tab of the model.
The results calculated by the model, based on this “business as usual” are surprisingly positive. They show the unfunded liability being consistently whittled away, such that by 2020 the consolidated systems are 72% funded, by 2025 they are 81% funded, by 2030 they are 91% funded, and by 2040 they are 118% funded.
All of that assumes, however, that from 2016 on, the consolidated system assets earn 7.5%. Every year without exception.
First Set of Downturn Cases: The markets fall 47% in 2017, and earn 5% per year thereafter.
This set of cases is to show the impact of a severe market correction, followed by modest returns from then on. This set of cases assumes a 7.5% return in 2016, followed by the crash in 2017, followed by modest 5% returns from then on. The five cases depicted here reflect escalating reductions to benefits. In all cases, what is solved for is the contribution – as a percent of payroll – required for the systems to achieve 100% funding by 2030. The baseline, by the way, under this earnings scenario – that is, if no changes are made to benefits or contributions in the face of this market correction – has the pension systems completely out of funds by 2041.
In the chart depicted below, five cases are presented. The first one makes no changes to benefits, solving for what contribution as a percent of payroll is necessary to restore 100% funding to the pension systems by 2050. As can be seen, the required contribution “$ Contr” is 80% of payroll. Recall that the consolidated average for California’s state/local pension systems is currently 32%, which equates to $34 billion per year. Put another way, it would cost California’s taxpayers another $51 billion per year if the market corrects by 47%, recovers at an earnings rate of 5% per year, and no changes are made to pension benefits.
The next three cases in the chart below consider changes to new hire benefits starting in 2020. Case #2 lowers the pension multiplier for new hires from 2.5% to 2.0%; a significant change that goes further than any pension reforms being seriously considered. As can be seen, it results in a lowering of the required contribution for these systems to restore 100% funding by 2050 to 66% of payroll, still more than twice what it is today.
Case #3 shows the impact of lowering the annual multiplier for new hires from 2.5% to 2.0%, combined with raising the eligible retirement age by five years. This results in a lowering of the required contribution for these systems to restore 100% funding by 2050 to 52% of payroll. Case 3 is highlighted because it probably represents the extreme limit of what current discussions indicate might eventually be negotiable.
In an attempt however to establish parameters for what additional reforms might yield, Case #4 shows the impact of lowering the pension multiplier for new hires still further, to 1.5%. The result is significant, bringing the required contribution for these systems to restore 100% funding by 2050 to 43% of payroll. Case #5 delves into territory that some would consider reasonable; it lowers the pension multiplier for existing employees to 2.0% per year – but only from now on, not retroactively. This yields significant savings, lowering the required contribution to 35% of payroll in order to achieve 100% funding by 2050. This scenario represents a viable way to preserve defined benefits without costing taxpayers tens of billions per year.
Second Set of Downturn Cases: The markets fall 47% in 2017, and earn 4% per year thereafter.
It is possible however that 5% returns per year after a severe market correction may be too much. After the last major stock market crash, in 1929, it took several decades before the cumulative average returns post-1929 reached 5% per year. The next set of cases examines the results of a 47% market downturn in 2017, followed by four percent returns from then on. The results are dramatically different from the 5% return scenario in the first set of cases, and illustrates the extraordinary sensitivity of pension fund solvency to the long-term rates of return.
In Case #6, just as in Case #1, no change has been made to the benefits either for new or existing employees. As can be seen, going from a 5% post crash annual average return to a 4% post crash annual average return causes the required contribution as a percent of payroll to jump from 80% (Case #1) to 113% (Case #6). This translates into an increase to taxpayers of $85 billion per year.
The next four cases in this 4% earnings scenario depict the same set of benefit reductions as in the previous 5% earnings scenario. But in this case, by the time you get to Case #10 (4% post crash earnings), which has the same set of reductions as Case #5 (5% post crash earnings), instead of whittling the required annual contribution (to achieve 100% funding by 2050) down to 35%, it is still at 48%. For this reason, additional cases are considered.
Cases #11 and #12 impose additional reductions to benefits on existing employees. By lowering the pension multiplier to 1.5% for existing employees – again, only from now on – the required contribution drops to 41%. By raising the age of eligibility for benefits by five years for existing employees, the required contribution drops to a manageable 37%.
PART THREE: CONCLUSION AND RECOMMENDATIONS
Based on historical trends, three key aggregated stock market ratios – price/earnings, price/sales, and price as a percent of GDP – all show that publicly traded U.S. stock are overvalued by approximately 50%. These indicators are supported by evidence that debt accumulation can no longer be useful as a tool to stimulate the economy or the stock markets, and that demographic trends – where an aging population is introducing more sellers into the markets – creates additional downward pressure on the stock markets.
It is easy enough to step back and claim that the rules have changed, that these unusually high stock market multiples can be sustained for additional decades, and that productivity improvements will enable the U.S. economy to support both massive debt and an aging population. Those who argue this position are betting that the U.S. economy, because of its diversity, sheer size, relatively lower levels of total market debt as a percent of GDP, relatively higher interest rates (i.e. still positive), stability and security, will become a refuge for wealth fleeing far more tumultuous economies elsewhere in the world. Staking the future of pension fund systems on this argument is a dangerous gamble. For more on why conservative rates-of-return may be inevitable, and imminent, read Michael Lebowitz’s two-part series, recently published by the California Policy Center, “The Death of the Virtuous Cycle,” and “The Fifteenth of August.”
For all of California’s pension funds consolidated, the analysis finds that if the market corrects downwards by 50%, then recovers to grow at 5% per year, without benefit reductions, the required annual contribution will rise to 80% of pension eligible payroll. This will cost California’s taxpayers an additional $50 billion per year. If the market corrects by 50% then recovers to grow at 4% per year, without benefit reductions, the required annual contribution will rise to 113% of pension eligible payroll. This scenario will cost California’s taxpayers an additional $86 billion per year. This analysis has also evaluated the impact of benefit reductions, exploring cases where only new employees are affected, as well as cases where benefits earned for future work by existing employees are also affected.
When evaluating what level of benefit reductions can preserve pension systems without increasing employer contributions, the study found the following:
In the case of a stock market negative correction of 50% followed by 5% annual growth, if the annual multiplier (the amount that is multiplied by years work times final salary to calculate an initial pension) is cut by 40% for new employees, and the annual retiree cost-of-living adjustment is cut by 40% for new employees, the retirement eligibility age is raised by five years for new employees, AND the annual multiplier is cut by 20% for existing employees for future work, the current level of payments would be sufficient to bring the systems to 100% funded status by 2050.
In the case of a stock market negative correction of 50% followed by 4% annual growth, if the annual multiplier (the amount that is multiplied by years work times final salary to calculate an initial pension) is cut by 40% for new employees AND existing employees for future work, and the annual retiree cost-of-living adjustment is cut by 40% for new employees AND existing employees, the retirement eligibility age is raised by five years for new employees AND existing employees, the current level of payments would be sufficient to bring the systems to 100% funded status by 2050.
Preserving defined benefit pensions in the face of a prolonged period of low investment returns will require the stakeholders in public employee pension funds to make hard choices. For example, lowering benefit accruals – just for future work by existing employees – can have a significant impact on reducing the required contribution. Reducing benefits for new employees is helpful, but the savings to be realized are decades in the future and are therefore heavily discounted when calculating a current pension system liability. For this reason, only restricting benefit reductions to new employees means they must be severe indeed. Balancing benefit reductions between existing employees (for future work) and new employees yields far less severe overall reductions – although as demonstrated, if the average market returns dip below 5% in the long-term – there is no incremental way to preserve system solvency without also dramatically increasing the required contributions.
Policymakers and stakeholders should plan now. They should negotiate, at the least, contingency plans to handle a major stock market correction that strikes an appropriate balance between costly tax increases and benefit reductions. Should there be a significant stock market downturn, taxpayers themselves will have depleted savings. They should not have to endure higher taxes to maintain public sector retirement accounts, when their own retirement accounts have been equally affected.
* * *
Ed Ring is the president of the California Policy Center.