Do changes in the economy affect the stock and bond markets?
From day to day, fluctuating prices in the financial markets are of great interest for high-frequency traders. But daily market movement is so volatile that economists call it a “random walk” whose next steps are unpredictable. To date, no clear pattern of daily movement has been found by researchers.
In the short term (over a span of months to a year), especially during transitional periods in the business cycle, economic conditions may appear to have a direct impact on the financial markets, and the financial markets, conversely, may appear to directly influence the economy. Around 2005, for instance, the housing bubble contributed to the overheated financial markets, and when the financial markets crashed in 2007-2008, the economy plunged into a severe recession. Clearly, over the short term, financial market performance and economic conditions were affected by each other.
Over a longer term (from years to decades), however, the relationship between financial market movements and the economy is unclear. In 2009, when the U.S. GDP contracted 2.8 percent, the Standard & Poor’s 500 index posted an annual return of 25.9 percent. The stock market regained its momentum long before the economy did. Moreover, when the economy was struggling with its slow recovery after 2009, the S&P 500 index averaged a positive annual gain of 15.7 percent from 2009 to 2014, rising beyond its pre-crisis level.
In this study we took a closer look at the long-term view—30 years—to see if any patterns were, in fact, discernible between the movements of the financial markets and the economy. We broke down the financial market into the two components—stocks and bonds—to see whether or how each moved relative to the broader economy. We used various economic indicators to measure the real economy, such as the money supply, consumer credit, manufacturers’ prices, unemployment claims, and so on.
We found that more than 99 percent of variation in major bond indexes correlated with the economy’s movements over the past 30 years, while less than one third of stock market fluctuations were explained by movements in the economy.
Because of the 30-year time frame of this study, the results will be of more interest for long-term investors—those who choose their portfolio and plan to stick to it for years rather than reacting to short-term economic fluctuations.
This is our first look at co-movements between the financial markets and the economy. We plan to expand our research in the future, choosing different periods to investigate how the co-movements vary over time.
Tracking the market, the economy, and their co-movements
We used long-term real data and state-of-the-art statistical methods to capture the movements in the bond market, the stock market, and the economy and determine the degree to which they moved together or were insulated from each other.
To sort this out, we needed to do two things. We needed to identify what we meant by bond market fluctuations, stock market fluctuations, and economic fluctuations. Second, we needed to find out how these market shifts and economic changes responded to one another.
In this analysis, we used the monthly growth rates of three different bond market indexes, three different stock market indexes, and six real economy variables from 1987 to 2014. From the data, we isolated four different cycles: the common cycle, the stock cycle, the bond cycle, and the business cycle.
The common cycle explains fluctuations shared by financial markets (stocks and bonds) and the economy over time, representing their interactions. The stock cycle accounts for stock market fluctuations that are shared by all stock indexes but are not shared by the economy or the bond market. The bond cycle covers bond market fluctuations that are shared by all bond indexes but are not shared by the economy or the stock market. The business cycle captures shared fluctuations among all economic indicators, and these fluctuations have no connection to the financial markets.
For each of the 12 variables (i.e., three stock indexes, three bond indexes, six economic variables), all fluctuations are explained by three unique factors: the common cycle, the group cycle (the stock cycle for stock market variables, the bond cycle for bond market variables, and the business cycle for economic variables), and the idiosyncratic factor (or everything we couldn’t explain).
We chose three stock indexes (S&P 500, NASDAQ 100, Russell 2000) to represent the stock market and three bond indexes (Barclays U.S. Government/Credit index, Barclays U.S. Treasury bond index, and Barclays U.S. Corporate High-Yield bond index) to represent the bond market. We chose six economic variables to represent the overall economy. These variables are: M2 money supply, the ISM index of manufacturers’ prices, consumer credit outstanding, new housing permits, initial claims for state unemployment insurance, and average workweek in manufacturing. The definition of each variable is presented in the table.
Financial market variables
Stock market variables
S&P 500 index includes 500 large-cap companies’ common stocks.
NASDAQ 100 index includes 100 of the largest non-financial companies listed on the NASDAQ.
Russell 2000 index includes 2000 small-cap to mid-cap companies.
Bond market variables
Barclays U.S. Government/Credit index includes both government and corporate debt issues that are investment grade (rated Baa/BBB or higher).
Barclays U.S. Treasury bond index: the public obligations of the U.S. Treasury.
Barclays U.S. Corporate High-Yield bond index includes USD-denominated, non-investment grade, fixed-rate, taxable corporate bonds.
M2 money supply includes notes and coins in circulation, traveler’s checks, checkable deposits, savings deposits, small-denomination time deposits, and retail money-market mutual fund shares.
ISM index of manufacturers’ prices: percentage of purchasing agents who report paying higher prices in the current month compared with the preceding month.
Consumer credit outstanding includes credit-card loans, auto loans, education loans, and so on, but no home mortgages or home-equity loans.
New housing permits: new private housing units authorized by local governments.
Initial claims for state unemployment insurance: the average number of persons who filed first-time claims for unemployment compensation each week that month.
Average workweek in manufacturing: the total of paid labor-hours of manufacturing production workers divided by the number of such workers for that week.
The analysis is based on a statistical modeling technique that is entirely data driven and where there are no behavioral assumptions imposed by the analyst. The findings were drawn from more than 14,000 iterations of a formula in which the results of these thousands of estimations converged in our final result. For a full description of the model and estimation techniques, see AIER’s working paper No. 003. (Liu, Jia. 2015. “Co-movements between the financial markets and the economy.” AIER WP-003, American Institute for Economic Research, Great Barrington, Mass.)
What we found: co-movement between the economy and bonds but not stocks
We estimated the four cycles and determined the importance of each cycle in explaining fluctuations in each of the 12 variables.
Charts 1-3 provide the results for all the variables in three groups: stock variables, bond variables, and economic variables. While only a small portion of the fluctuations in the stock market can be accounted for by fluctuations in the overall economy, Chart 1 illustrates that there are differences among our three stock market indexes. The relationship with the overall economy—measured by the common cycle—is strongest for the S&P 500 at just over 30 percent and weakest for the Russell 2000 at just over 1 percent. The different performance of the S&P 500, the NASDAQ 100 and the Russell 2000 is not surprising. In fact, it is because their performances are different that we need all three indexes to capture the stock market as a whole. Our interpretation for this difference is that the S&P 500 captures diversity across the economy; the NASDAQ 100 is more heavily weighted toward tech stocks and large-cap companies, and the Russell 2000 is more heavily weighted toward smaller companies.
While the common cycle only accounts for a small portion of the variation in the stock market for the past 30 years, the stock cycle accounts for overall about 40 percent of the long-term variations in the stock market. However, there is still a big portion of variation in stock indexes that remains unaccounted for by either the common cycle or the stock cycle, reflecting the individual variability of each stock index and the high uncertainty and high risk of investing in stocks.
In the bond market, over 99 percent of variations in the U.S. government/credit index and the U.S. Treasury bond index are explained by the same cycle—the common cycle (see Chart 2). This has two implications. First, the movements of the U.S. government/credit index and the U.S. Treasury bond index are nearly identical over the long term. These two indexes will not provide meaningful diversification for investors. Second, both of these bond indexes move in relation to the broader economy, so broader economic conditions are important to consider for bond market performance.
Among the three bond indexes included in this study, the U.S. corporate high-yield (low-grade) bond index has the weakest correlation to the common cycle, with 55 percent of its variation captured by the common cycle (see Chart 2). Therefore, including this index along with one of the other bond indexes would provide diversification and reduce portfolio risk.
In the economic landscape, only two of the six economic variables show a strong relationship with the common cycle (see Chart 3). They are consumer credit and the money supply, with 91 percent and 80 percent of variation accounted for, respectively. Other economic variables have little variation due to the common cycle. Recall that almost all of the movement in the U.S. government/credit bond index and the Treasury bond index is explained by the common cycle, which implies that the two bond indexes are highly correlated with the money and credit variables.
What our study means for investors
We found little evidence of long-term correlation between stocks and economic conditions. The results of this study point to the difficulty—if not impossibility—of using broader economic data to make long-term stock market decisions.
Knowing the risk levels of financial instruments is critical when you are constructing a portfolio. In this study we used three stock indexes and found that about half of the variation of two of these—the NASDAQ 100 and the Russell 2000—remain unexplained, indicating high uncertainty and high risk.
In addition, we found that the three stock indexes had different profiles with respect to movement with our cycles. We take this to mean that holding stocks from different stock indexes provides an important degree of portfolio diversification.
For bond investors, bond returns, especially for government bonds and high grade (low-yield) corporate bonds, proved to be highly correlated with the money supply and consumer credit in the past three decades. If this relationship holds in the future, it means that bond investors can use trends in money and credit indicators to inform their bond investment decisions. Also, government bonds and high-grade corporate bonds appear to move together, so including both types of bonds in portfolios may not provide sufficient diversification.
To sum up, there are five takeaways from the above analysis:
Overall, the stock indexes show a much weaker connection with the overall economy than do bond indexes. The long-term trend in economic indicators fails to provide a clear perspective of stock market performance.
The stock indexes turn out to be correlated with one another on some level. But a big portion of variation in each stock index remains unexplained, meaning high risk of investing in stocks.
The bond market shows a high correlation with economic conditions, especially money and credit indicators.
Government bonds and high-grade (low-yield) corporate bonds move closely with each other, so including both rather safe bonds in a portfolio may not provide sufficient diversification.
Corporate low-grade (high-yield or high-risk) bonds only partially move the way government bonds and high-grade (low-yield) corporate bonds do, so including corporate low-grade bonds with rather safe government bonds or high-grade corporate bonds may reduce portfolio risk.
The current study aimed to investigate how financial markets related to the economy over the past 30 years. The results focused on the big picture, providing an overall perspective of the relationship between financial markets and economic conditions in the long run.
But some would be interested in how the relationship has changed during different periods. For instance, we found that the stock market barely moved in correlation with the economy over the past decades. But during the financial crisis of 2008, financial markets and the economy both faced a meltdown, moving down at the same time. Therefore, the absence of a relationship between the stock market and the economy over the long term very likely doesn’t apply to transitional periods or episodic shocks.
In the future, we will shift the focus from the long-term to the short-term relationship between financial markets and the economy, which should be of interest for short-term investors who like to adjust their investment portfolios in response to significant or abrupt changes in economic conditions.