The authors first summarize the previous research; some of which was in support of money supply changes causing stock prices, and others rejecting that notion in support of the Efficient Market Hypothesis. They also note more recent research that showed unanticipated changes in money supply caused a drop in stock prices; they cite several papers that attribute this effect to increased money supply causing investors to expect increased inflation, which depresses stock prices for a number of reasons (e.g., decreased real after-tax earnings, the opportunity cost of owning inflation linked assets like real estate and commodities, mistakes in comparing real returns to nominal returns, or Federal Reserve reactions and counter-measures such as raising interest rates).
The authors created a regression that included the change in stock prices as the dependent variable, and the actual minus expected change in the M1 as the independent variable. To determine the expected change, they used the median expectation of about 60 money market participants in the weekly survey by Money Market Services. They found that an unexpected increase in money supply of $1 billion usually caused a 0.7 fall in the stock price. The regressions did, however, exhibit relatively small R^2s.
The authors consider several other questions related to this. In particular, they find that stock price reactions to money supply changes become more volatile after the Fed beings its reserve-aggregate approach to monetary control; the stock price reactions to unanticipated changes in money supply is immediate; and stock prices do not move in anticipation of unexpected money supply changes.
PEARCE, D. K., & ROLEY, V. V. (1983). The Reaction of Stock Prices to Unanticipated Changes in Money: A Note. Journal of Finance, 38(4), 1323–1333.
Tuesday, October 30, 2018
Sunday, October 28, 2018
Rational Expectations and the Impact of Money upon Stock Prices
The author notes the prior studies, noting significant relationships between money supply and stock prices; they also note that most recent studies have concluded that the stock market anticipates changes in the money supply and not the other way around. Their addition to the area is by using the Barro equation, but splitting future changes in money supply into anticipated and unanticipated changes. Under the efficient market hypothesis, unanticipated changes in the money supply would abruptly affect stock prices; however, expected changes would not.
He first uses the Barro equation to produce a regression with M2 as the dependent variable, and lagged money supply figures, the unemployment rate, and federal expenditures as the independent variables. He shows these variable to explain 94% of the changes in M2 and to all be significant at a 95% confidence level. In the second part, the author produces a regression with stock returns as the dependent variable; for the independent variables, he uses the previous money supply equation as the ANTICIPATED money supply changes, and he uses the residuals from this equation as the UNANTICIPATED money supply changes.
The author found that the unanticipated changes in money supply were significant in changing stock returns; and the anticipated changes in money supply were insignificant. He then produces the same equation with further lags, and the significance is even more pronounced in support of stock prices anticipating money supply changes. He proposes that the link between money supply and stock returns found by other studies may in fact be a link between unanticipated changes in money supply and the act of sophisticated investors closing that gap.
Sorensen, E. H. (1982). Rational Expectations and the Impact of Money upon Stock Prices. Journal of Financial & Quantitative Analysis, 17(5), 649–662.
He first uses the Barro equation to produce a regression with M2 as the dependent variable, and lagged money supply figures, the unemployment rate, and federal expenditures as the independent variables. He shows these variable to explain 94% of the changes in M2 and to all be significant at a 95% confidence level. In the second part, the author produces a regression with stock returns as the dependent variable; for the independent variables, he uses the previous money supply equation as the ANTICIPATED money supply changes, and he uses the residuals from this equation as the UNANTICIPATED money supply changes.
The author found that the unanticipated changes in money supply were significant in changing stock returns; and the anticipated changes in money supply were insignificant. He then produces the same equation with further lags, and the significance is even more pronounced in support of stock prices anticipating money supply changes. He proposes that the link between money supply and stock returns found by other studies may in fact be a link between unanticipated changes in money supply and the act of sophisticated investors closing that gap.
Sorensen, E. H. (1982). Rational Expectations and the Impact of Money upon Stock Prices. Journal of Financial & Quantitative Analysis, 17(5), 649–662.
Saturday, October 27, 2018
Portfolio - October 27, 2018
| Company Name | Ticker | Weight |
| Intel Corporation | INTC | 10.97% |
| Total SA | TOT | 8.60% |
| Mondelez International Inc | MDLZ | 7.69% |
| Kellogg Company | K | 6.52% |
| Exxon Mobil Corp | XOM | 6.26% |
| FedEx Corp | FDX | 6.10% |
| Chevron Corporation | CVX | 5.69% |
| Microsoft Corporation | MSFT | 5.59% |
| Comcast Corporation | CMCSA | 4.98% |
| Procter and Gamble Co | PG | 4.05% |
| Walt Disney Co | DIS | 3.85% |
| Motorola Solutions Inc | MSI | 3.54% |
| Parker Hannifin Corp | PH | 3.49% |
| Pfizer Inc | PFE | 3.11% |
| Discover Financial Services | DFS | 3.05% |
| JP Morgan Chase and Co | JPM | 2.74% |
| Cognizant Technology Solutions Corp | CTSH | 2.61% |
| Lockheed Martin Corp | LMT | 2.04% |
| Phillips 66 | PSX | 1.87% |
| Verizon Communications Inc | VZ | 1.16% |
| Nike Inc | NKE | 1.06% |
| Simon Property Group Inc | SPG | 1.05% |
| West Pharmaceutical Services Inc | WST | 0.99% |
| Union Pacific Corp | UNP | 0.96% |
| Avalonbay Communities Inc | AVB | 0.91% |
| Archer Daniels Midland Company | ADM | 0.65% |
| Sempra Energy | SRE | 0.41% |
| Lowes Companies Inc | LOW | 0.06% |
| 100.00% |
Friday, October 26, 2018
Determinants of Common Stock Prices: A Time Series Analysis
The authors begin with a quick history of the studies around money supply and stock prices, noting that some researches have confirmed an unidirectional causation while others have rebuffed it. At stake is whether the efficient market hypothesis holds (i.e., that only current and future estimates of money supply levels dictate stock prices); if future stock prices are caused by changes in the money supply, then one would reject the efficient market hypothesis and abnormal trading profits could be earned.
The authors use several determinants of stock prices (e.g., money supply, rate of change in money supply, corporate interest rate, and a measure of risk) to see if there is a causal relationship between them and stock prices. They form a set of 4 equations for each determinant of stock prices. The 1st equation includes current and lagged values of the determinant as independent variables; the 2nd equation includes, current, lagged, and future values of the determinant as independent variables; and the 3rd and 4th equations reverse the dependent and independent variables of equations 1 and 2.
Through these equations, the authors determine that there is a statistical relationship between the determinants and stock prices. However, they find that there is no significant causality relationship running from money supply to stock prices; in fact, they find that current stock prices have an influence on future money supply. Which is consistent with the efficient market hypothesis, whereby, future changes are anticipated by stock price movements.
KRAFT, J., & KRAFT, A. (1977). Determinants of Common Stock Prices: A Time Series Analysis. Journal of Finance, 32(2), 417–425.
The authors use several determinants of stock prices (e.g., money supply, rate of change in money supply, corporate interest rate, and a measure of risk) to see if there is a causal relationship between them and stock prices. They form a set of 4 equations for each determinant of stock prices. The 1st equation includes current and lagged values of the determinant as independent variables; the 2nd equation includes, current, lagged, and future values of the determinant as independent variables; and the 3rd and 4th equations reverse the dependent and independent variables of equations 1 and 2.
Through these equations, the authors determine that there is a statistical relationship between the determinants and stock prices. However, they find that there is no significant causality relationship running from money supply to stock prices; in fact, they find that current stock prices have an influence on future money supply. Which is consistent with the efficient market hypothesis, whereby, future changes are anticipated by stock price movements.
KRAFT, J., & KRAFT, A. (1977). Determinants of Common Stock Prices: A Time Series Analysis. Journal of Finance, 32(2), 417–425.
Wednesday, October 24, 2018
Money and Stock Prices: Market Efficiency and the Lag in Effect of Monetary Policy
The author starts with a review of the literature. In particular, the "Predictive Monetary Portfolio" explanation theorizes how more money supply causes an imbalance of money in investors' portfolios, causing them to trade money for stocks or other assets, causing a lagged increase in stock prices. However, the efficient market hypothesis would claim there should be no lag and no ability for an investor to make excess profits using this lag; several studies (i.e., Sprinkel and Homa/Jaffee) have seemed to contradict the EMH in this regard.
The author suggests that prior studies (in particular Cooper 1972) have been faulty in regard to the EMH in that they don't distinguish between whether money supply affects stock prices (which is agreeable with the EMH) or money supply affects stock prices with a lag (which would go against the EMH). The author proposes a "Non-Predictive" Monetary Portfolio model that explains how current and anticipated future money supply changes affect current stock prices, which would be in line with the EMH.
The author then goes into ways that an investor may earn excess returns that do not necessarily counter the EMH. For example, he may have a superior predictive model unavailable to other investors using publicly available data, or he may have access to superior data that is costly to obtain. Also, certain trading rules may appear to provide excess returns, but may in fact have performance calculation flaws (e.g., the ineffective treatment of dividends in short sales, neglect of brokerage and other costs uncharacteristic of a buy-and-hold strategy, only showing returns for the in-sample period and neglecting an out-of-sample period, failure to adjust for risk, use of unattainable prices for the transactions, using data that would be unavailable at the time of trading decision).
The author then proposes a regression to test whether the growth in money supply explains stock returns. The dependent variable is stock returns, and the independent variables in the equation are the current growth rate in money supply and several lagged growth rates in the money supply. The thought is that if the lagged variables turn out to be significant, then there is evidence to rebuff the EMH and to conclude that trading rules can be made to exploit the disconnect; however, if the contemporaneous variable is significant and the lagged variables are not, then there is evidence to claim the market is efficient and there is no way to abnormally profit from the discrepancy. The authors caution the researcher to use money supply data as of or after the publishing data (not the as of date), in order to prevent using data that was unavailable at the time of decision making.
The authors find in their regression, that over the entire sample period of August 1948 - March 1970, the independent variables were determined to be insignificant at a 5% level, and the adjusted R^2 was below 0.04; both of which would support the EMH and the inability to earn excess returns from lagged monetary supply variables. They then test their non-predictive MP regression and find slightly significant support for contemporaneous explanation between money supply changes and stock returns. He then uses a regression to anticipate the non-predictive MP model that takes into account contemporaneous and future anticipated money supply levels; he finds the independent variables to be significant at a 1% level and the R^2 to be sufficiently high to err in favor of the EMH.
The author then does a trading rule simulation by taking the regression equation for the 1/47 - 9/56 period, and using it to determine whether to invest in stocks or commercial paper depending on whether the regression's estimate of the stock returns exceed the commercial paper return. The backtesting shows that when incorporating the publication lag of the monetary data into the equation, the trading rules do not beat a buy-and-hold strategy; however, not incorporating the publication lag produces extremely high returns compared to the buy-and-hold strategy. This result emphasizes the importance of timing the data appropriately in the model; otherwise, a false interpretation could be construed. The result is that if someone knew what the money supply figures would be before they are published, they could profit from a strategy using money supply as an independent variable.
The author then goes into exploring prior studies. First, he bashes Sprinkel's landmark study by explaining how Sprinkel backed into his trading strategy ex post, which would be impossible for an investor to do ex ante. Palmer's study (which followed Sprinkel's) also gets a good thrashing due to his use of a moving average of stock returns, which would be correlated with the lagged variables causing misleading results. He then pointed out how Reilly and Lewis found evidence to support the EM, but continued to naively support the Sprinkel study. He then goes on to rebuff Keran's, Hamburger and Kochen's, Homa and Jaffee's, and Cooper's studies citing biases in choices and use of unavailable data or neglecting to use appropriate tests to backup their conclusions.
Rozeff, M. S. (1974). MONEY AND STOCK PRICES: Market efficiency and the lag in effect of monetary policy. Journal of Financial Economics, 1(3), 245–302.
The author suggests that prior studies (in particular Cooper 1972) have been faulty in regard to the EMH in that they don't distinguish between whether money supply affects stock prices (which is agreeable with the EMH) or money supply affects stock prices with a lag (which would go against the EMH). The author proposes a "Non-Predictive" Monetary Portfolio model that explains how current and anticipated future money supply changes affect current stock prices, which would be in line with the EMH.
The author then goes into ways that an investor may earn excess returns that do not necessarily counter the EMH. For example, he may have a superior predictive model unavailable to other investors using publicly available data, or he may have access to superior data that is costly to obtain. Also, certain trading rules may appear to provide excess returns, but may in fact have performance calculation flaws (e.g., the ineffective treatment of dividends in short sales, neglect of brokerage and other costs uncharacteristic of a buy-and-hold strategy, only showing returns for the in-sample period and neglecting an out-of-sample period, failure to adjust for risk, use of unattainable prices for the transactions, using data that would be unavailable at the time of trading decision).
The author then proposes a regression to test whether the growth in money supply explains stock returns. The dependent variable is stock returns, and the independent variables in the equation are the current growth rate in money supply and several lagged growth rates in the money supply. The thought is that if the lagged variables turn out to be significant, then there is evidence to rebuff the EMH and to conclude that trading rules can be made to exploit the disconnect; however, if the contemporaneous variable is significant and the lagged variables are not, then there is evidence to claim the market is efficient and there is no way to abnormally profit from the discrepancy. The authors caution the researcher to use money supply data as of or after the publishing data (not the as of date), in order to prevent using data that was unavailable at the time of decision making.
The authors find in their regression, that over the entire sample period of August 1948 - March 1970, the independent variables were determined to be insignificant at a 5% level, and the adjusted R^2 was below 0.04; both of which would support the EMH and the inability to earn excess returns from lagged monetary supply variables. They then test their non-predictive MP regression and find slightly significant support for contemporaneous explanation between money supply changes and stock returns. He then uses a regression to anticipate the non-predictive MP model that takes into account contemporaneous and future anticipated money supply levels; he finds the independent variables to be significant at a 1% level and the R^2 to be sufficiently high to err in favor of the EMH.
The author then does a trading rule simulation by taking the regression equation for the 1/47 - 9/56 period, and using it to determine whether to invest in stocks or commercial paper depending on whether the regression's estimate of the stock returns exceed the commercial paper return. The backtesting shows that when incorporating the publication lag of the monetary data into the equation, the trading rules do not beat a buy-and-hold strategy; however, not incorporating the publication lag produces extremely high returns compared to the buy-and-hold strategy. This result emphasizes the importance of timing the data appropriately in the model; otherwise, a false interpretation could be construed. The result is that if someone knew what the money supply figures would be before they are published, they could profit from a strategy using money supply as an independent variable.
The author then goes into exploring prior studies. First, he bashes Sprinkel's landmark study by explaining how Sprinkel backed into his trading strategy ex post, which would be impossible for an investor to do ex ante. Palmer's study (which followed Sprinkel's) also gets a good thrashing due to his use of a moving average of stock returns, which would be correlated with the lagged variables causing misleading results. He then pointed out how Reilly and Lewis found evidence to support the EM, but continued to naively support the Sprinkel study. He then goes on to rebuff Keran's, Hamburger and Kochen's, Homa and Jaffee's, and Cooper's studies citing biases in choices and use of unavailable data or neglecting to use appropriate tests to backup their conclusions.
Rozeff, M. S. (1974). MONEY AND STOCK PRICES: Market efficiency and the lag in effect of monetary policy. Journal of Financial Economics, 1(3), 245–302.
Saturday, October 20, 2018
Money and Stock Prices: The Channels of Influence
The authors note several flaws with prior studies regarding the money supply and stock prices. In particular, most of the recent studies only use the period after the Korean war, the studies only find an indirect relationship (e.g., through changes in interest rates or earnings), and they generally neglect the risk aspect which has been a tenant of investment research.
The authors go into explaining how the change in money supply affects all the components of the earnings discount model (i.e., the risk-free yield, earnings expectations, and the risk premium); ***note that Homa/Jaffee (1971) used a dividend discount model.*** The authors first discuss the "liquidity effect", whereby more money supply reduces the risk-free rate; they suggest that stock prices may be more responsive, in the short run, to changes in the money supply than bonds. Next is the "earnings effect", whereby increases in the money supply create more demand for goods/services which translates into greater corporate profits; they suggest that changes in the money supply precede changes in stock prices which precede changes in earnings. Finally, the authors discuss the "risk premium effect", whereby greater volatility of money supply levels may cause greater volatility in the economy requiring an increase in the risk premium resulting in a reduction in stock prices.
The authors perform two studies. In the first study (the "partial effect"), the authors follow a paper by Keran (1971) to investigate how the money supply affects stock prices through an econometric model. Keran had suggested through his results that money supply only has an indirect effect on stock prices; however, the authors split several of the variables into their components (e.g., split interest rates into real and inflation components, etc) and find that money supply has BOTH an indirect and direct effect on stock prices. That is, the money supply growth variable is still significant after making the regression more granular. They find that the inflation relationship is now insignificant and the money supply changes are now a significant determinant of stock price levels.
In the second study (the "total effect"), the authors make the independent variables be lagged money growth rate going back in 9 successive quarters; the dependent variable is either the risk free rate, corporate bond rate, S&P stock index, or dividend yield in 4 separate equations. They find that the equation with the stock price level as the dependent variable has a larger R^2 than Keran's regression in the "partial effect". The authors use this as further evidence that the money supply has a direct effect on stock prices (not just indirectly through interest rates), contradicting the conventional wisdom.
The authors then go on to regress an equation using data back to 1871 in an effort to support the volatility of the money supply causing higher risk premiums.
HAMBURGER, M. J., & KOCHIN, L. A. (1972). Money and Stock Prices: The Channels of Influence. Journal of Finance, 27(2), 231–249.
The authors go into explaining how the change in money supply affects all the components of the earnings discount model (i.e., the risk-free yield, earnings expectations, and the risk premium); ***note that Homa/Jaffee (1971) used a dividend discount model.*** The authors first discuss the "liquidity effect", whereby more money supply reduces the risk-free rate; they suggest that stock prices may be more responsive, in the short run, to changes in the money supply than bonds. Next is the "earnings effect", whereby increases in the money supply create more demand for goods/services which translates into greater corporate profits; they suggest that changes in the money supply precede changes in stock prices which precede changes in earnings. Finally, the authors discuss the "risk premium effect", whereby greater volatility of money supply levels may cause greater volatility in the economy requiring an increase in the risk premium resulting in a reduction in stock prices.
The authors perform two studies. In the first study (the "partial effect"), the authors follow a paper by Keran (1971) to investigate how the money supply affects stock prices through an econometric model. Keran had suggested through his results that money supply only has an indirect effect on stock prices; however, the authors split several of the variables into their components (e.g., split interest rates into real and inflation components, etc) and find that money supply has BOTH an indirect and direct effect on stock prices. That is, the money supply growth variable is still significant after making the regression more granular. They find that the inflation relationship is now insignificant and the money supply changes are now a significant determinant of stock price levels.
In the second study (the "total effect"), the authors make the independent variables be lagged money growth rate going back in 9 successive quarters; the dependent variable is either the risk free rate, corporate bond rate, S&P stock index, or dividend yield in 4 separate equations. They find that the equation with the stock price level as the dependent variable has a larger R^2 than Keran's regression in the "partial effect". The authors use this as further evidence that the money supply has a direct effect on stock prices (not just indirectly through interest rates), contradicting the conventional wisdom.
The authors then go on to regress an equation using data back to 1871 in an effort to support the volatility of the money supply causing higher risk premiums.
HAMBURGER, M. J., & KOCHIN, L. A. (1972). Money and Stock Prices: The Channels of Influence. Journal of Finance, 27(2), 231–249.
Portfolio - October 10, 2018
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Friday, October 19, 2018
The Supply of Money and Common Stock Prices
The authors begin by suggesting a theoretical explanation for why changes in money supply and common stocks may be related. They review the Dividend Discount Model, noting that investors would value a stock based on the rate and growth of dividends, the risk free rate, and an equity risk premium. Their study attempts to show that changes in the money supply are positively related to the rate and growth of dividends (which increase a stock's value), and negatively related to the risk free rate and equity risk premium (which decrease a stock's value). The authors suggest that a decrease in the money supply will cause an
increase in interest rates, which would cause a decrease in
expenditures, which will decrease sales, which will ultimately decrease
profits and dividends.
The authors study the period January 1954 - April 1969 using the S&P 500 and M1. They develop a regression equation that produces an R^2 of 0.968, but note a significant amount of serial correlation. They then use a series of regressions to forecast 1-month ahead out-of-sample (1960-1969) stock prices based on money supply changes; in doing so, their regressions explain approximately 90% of the error, meaning a very good fit.
The authors then used a regression equation using factors of unemployment rate, inflation rate, and level of the US international reserve position in order to predict the growth in money supply, under the assumption that the Federal Reserve would makes its decisions based on the levels of these factors. They find their regression reasonably explains the growth rate of money supply, but note that the errors may stem from the Federal Reserve's changes in policy throughout different regimes during the sample period.
The authors then use simulations to test the level of stock price forecasting regression under normal investment decision making conditions. They give an investor 2 options to invest in: S&P 500 or 3-month treasury bills. Based on the prediction of the regression equation, the investor chooses to go all-in on stocks or treasuries. There are 6 simulations performed; 3 with margin and 3 without; and the 3 assuming perfect foresight, naive extrapolation, or regression prediction of the money supply variables that go into the regression equation.
The results of the market test show that an investor could have used this regression to beat a buy-and hold strategy by 2% CAY on return and slightly reduce risk; although, this could have only been done by reasonably predicting (either through perfect foresight or regression) the money supply variables that go into the regression equation. Naive extrapolation of the money supply variable underperformed the buy-and-hold strategy.
HOMA, K. E., & JAFFEE, D. M. (1971). The Supply of Money and Common Stock Prices. Journal of Finance, 26(5), 1045–1066.
The authors study the period January 1954 - April 1969 using the S&P 500 and M1. They develop a regression equation that produces an R^2 of 0.968, but note a significant amount of serial correlation. They then use a series of regressions to forecast 1-month ahead out-of-sample (1960-1969) stock prices based on money supply changes; in doing so, their regressions explain approximately 90% of the error, meaning a very good fit.
The authors then used a regression equation using factors of unemployment rate, inflation rate, and level of the US international reserve position in order to predict the growth in money supply, under the assumption that the Federal Reserve would makes its decisions based on the levels of these factors. They find their regression reasonably explains the growth rate of money supply, but note that the errors may stem from the Federal Reserve's changes in policy throughout different regimes during the sample period.
The authors then use simulations to test the level of stock price forecasting regression under normal investment decision making conditions. They give an investor 2 options to invest in: S&P 500 or 3-month treasury bills. Based on the prediction of the regression equation, the investor chooses to go all-in on stocks or treasuries. There are 6 simulations performed; 3 with margin and 3 without; and the 3 assuming perfect foresight, naive extrapolation, or regression prediction of the money supply variables that go into the regression equation.
The results of the market test show that an investor could have used this regression to beat a buy-and hold strategy by 2% CAY on return and slightly reduce risk; although, this could have only been done by reasonably predicting (either through perfect foresight or regression) the money supply variables that go into the regression equation. Naive extrapolation of the money supply variable underperformed the buy-and-hold strategy.
HOMA, K. E., & JAFFEE, D. M. (1971). The Supply of Money and Common Stock Prices. Journal of Finance, 26(5), 1045–1066.
Wednesday, October 17, 2018
Alphas in disguise: A new approach to uncovering them
The Carhart 4-factor model has become the academic standard for explaining returns; however, recently passive indices have been shown to have non-zero alphas in their 4-factor models, which should not be so for passive indices that represent the market. Either the wrong factors are being selected or there are errors in the calculations of those factors used.
Previous studies have suggested that one error may be the SMB calculation: some small funds may actually have big companies in them, because, since big companies are small relative to the whole market, causing a biased positive SMB estimate. This goes for the other winner/loser, value/growth factors as well. To mitigate that error, other studies have reduced alpha in indices by including only only US Equities in the market portfolio, make the HML and SMB portfolios be value-weighted instead of equal-weighted, separate value/growth stocks and small/big stocks, and make the sizes more granular (such as large-cap, mid-cap, small-cap, etc.).
The authors study 1,383 US equity mutual funds over the period January 1992 to October 2013, who designate the S&P 500 as their benchmark. They then make adjustments to the Carhart 4-factors (obtained from Ken French's website) that will bring the alpha to zero. In their calculations, they find both active and passive funds perform worse on an adjusted alpha basis than what is implied by the unadjusted Carhart 4-factor model.
Also noted is that the passive tracker funds performed the worst over the period, showing significant negative alphas; which is counter-intuitive, considering a passive tracker fund should have a zero alpha. Active mutual funds perform outperform in some periods and under-perform in others.
Chinthalapati, V., Mateus, C., & Todorovic, N. (2017). Alphas in disguise: A new approach to uncovering them. International Journal of Finance & Economics, 22(3), 234–243.
Previous studies have suggested that one error may be the SMB calculation: some small funds may actually have big companies in them, because, since big companies are small relative to the whole market, causing a biased positive SMB estimate. This goes for the other winner/loser, value/growth factors as well. To mitigate that error, other studies have reduced alpha in indices by including only only US Equities in the market portfolio, make the HML and SMB portfolios be value-weighted instead of equal-weighted, separate value/growth stocks and small/big stocks, and make the sizes more granular (such as large-cap, mid-cap, small-cap, etc.).
The authors study 1,383 US equity mutual funds over the period January 1992 to October 2013, who designate the S&P 500 as their benchmark. They then make adjustments to the Carhart 4-factors (obtained from Ken French's website) that will bring the alpha to zero. In their calculations, they find both active and passive funds perform worse on an adjusted alpha basis than what is implied by the unadjusted Carhart 4-factor model.
Also noted is that the passive tracker funds performed the worst over the period, showing significant negative alphas; which is counter-intuitive, considering a passive tracker fund should have a zero alpha. Active mutual funds perform outperform in some periods and under-perform in others.
Chinthalapati, V., Mateus, C., & Todorovic, N. (2017). Alphas in disguise: A new approach to uncovering them. International Journal of Finance & Economics, 22(3), 234–243.
Tuesday, October 16, 2018
Stock prices, money supply, and interest rates: the question of causality
The question of causality between money supply and stock prices has been hotly debated. Some researchers hypothesize that changes in the money supply cause changes in the economy that require investors to alter their holdings of stocks. However, other researchers hypothesize that anticipated changes in stock prices are a front-runner to changes in the economy; and an improving stock market may mean the economy is improving and will result in an increase in the money supply to full-fill loan demand.
The authors study the period January 1980 - July 1986 using weekly S&P 500 and M1 data. Their findings from their Granger-Sims tests show that causality runs from money supply to stock prices, then stock prices to money supply. Through their regressions, they find that trying to predict stock prices using past money supply changes only is not very useful.
Hashemzadeh, N., & Taylor, P. (1988). Stock prices, money supply, and interest rates: the question of causality. Applied Economics, 20(12), 1603.
The authors study the period January 1980 - July 1986 using weekly S&P 500 and M1 data. Their findings from their Granger-Sims tests show that causality runs from money supply to stock prices, then stock prices to money supply. Through their regressions, they find that trying to predict stock prices using past money supply changes only is not very useful.
Hashemzadeh, N., & Taylor, P. (1988). Stock prices, money supply, and interest rates: the question of causality. Applied Economics, 20(12), 1603.
Monday, October 15, 2018
M1, M2, and the U.S. Equity Exchanges
The link between money supply and equity returns is well established. In the 70s, the focus was establishing a link between the two. In the 80s and 90s, there were many studies exploring the causality of the link; and during that time there was a conflict of whether changes in the money supply caused changes in stock prices, or whether it was actually the other way around. Also, in nearly all studies, M1 is usually the measure of money supply used. The authors, however, suggest there is a gap in the literature related
to M2's predictive power, and they explore this relationship as well as settle the causality issue.
The authors use the S&P500, Dow Jones Industrial Average, Nasdaq, and Wilshire 5000 as the equity components, and the M1 and M2 as money supply measures over the period January 1984 to November 2010. Their study finds the M1 can cause changes in stock prices, but only over long periods of time; the M2 measure, however, can cause changes in stock prices over very short periods of time, and is shown empirically to be a better predictor stock returns. The authors also note that the Fed's policy of quantitative easing over the period of their writing is in line with these findings.
Parhizgari, A. M., & Nguyen, D. (2011). M1, M2, and the U.S. Equity Exchanges. Frontiers in Finance & Economics, 8(2), 112–135.
The authors use the S&P500, Dow Jones Industrial Average, Nasdaq, and Wilshire 5000 as the equity components, and the M1 and M2 as money supply measures over the period January 1984 to November 2010. Their study finds the M1 can cause changes in stock prices, but only over long periods of time; the M2 measure, however, can cause changes in stock prices over very short periods of time, and is shown empirically to be a better predictor stock returns. The authors also note that the Fed's policy of quantitative easing over the period of their writing is in line with these findings.
Parhizgari, A. M., & Nguyen, D. (2011). M1, M2, and the U.S. Equity Exchanges. Frontiers in Finance & Economics, 8(2), 112–135.
Sunday, October 14, 2018
Money Supply and Stock Prices: A Probabilistic Approach
A relationship between money supply and stock prices is fairly recognized in the literature; Sprinkel and Palmer have tried to determine whether the money supply can predict stock prices. If this could be done, an investor could allocate his capital to and from the market portfolio, or in and out of high and low beta stocks, in an attempt to time the market.
The authors use a probability function to predict when a turning point in the money supply yields a turning point in stock prices; and they boil this down to the combination (i.e., an "efficiency index") of a "reliability index" (i.e., what ratio of predicted turning points were true turning points) and a "opportunity loss index" (i.e., what ratio of true turning points were predicted by the system).
They then used M1 (currency held + demand deposits) and M2 (M1 + time deposits) components of the money supply and the Standard and Poor's 425 over the period January 1948 to December 1970 in their analysis. Their results show that using a 3% filter for M2 seems to provide the best reliability index; wherein, the trough signals using M2 are 65% accurate and the peak signals using M2 are 59% accurate. However, using the higher filter of 3% yielded a lower opportunity loss index, meaning there were several true turning points undetected by the system.
Overall, their efficiency index finds that the 2% M1/5% Stocks filter works best for peaks, and the 1% M2/5% Stocks filter works best at troughs; although the 1% M1/5% Stocks is close to the 1% M2/5% Stocks for troughs. As such, using M1 with relatively sensitive filters seems to provide the best balance of reliability and opportunity loss for predicting turning points in stocks.
Gupta, M. C. (1974). Money Supply and Stock Prices: A Probabilistic Approach. Journal of Financial & Quantitative Analysis, 9(1), 57–68.
The authors use a probability function to predict when a turning point in the money supply yields a turning point in stock prices; and they boil this down to the combination (i.e., an "efficiency index") of a "reliability index" (i.e., what ratio of predicted turning points were true turning points) and a "opportunity loss index" (i.e., what ratio of true turning points were predicted by the system).
They then used M1 (currency held + demand deposits) and M2 (M1 + time deposits) components of the money supply and the Standard and Poor's 425 over the period January 1948 to December 1970 in their analysis. Their results show that using a 3% filter for M2 seems to provide the best reliability index; wherein, the trough signals using M2 are 65% accurate and the peak signals using M2 are 59% accurate. However, using the higher filter of 3% yielded a lower opportunity loss index, meaning there were several true turning points undetected by the system.
Overall, their efficiency index finds that the 2% M1/5% Stocks filter works best for peaks, and the 1% M2/5% Stocks filter works best at troughs; although the 1% M1/5% Stocks is close to the 1% M2/5% Stocks for troughs. As such, using M1 with relatively sensitive filters seems to provide the best balance of reliability and opportunity loss for predicting turning points in stocks.
Gupta, M. C. (1974). Money Supply and Stock Prices: A Probabilistic Approach. Journal of Financial & Quantitative Analysis, 9(1), 57–68.
Saturday, October 13, 2018
Portfolio - October 13, 2018
| Company Name | Ticker | Weight |
| Intel Corporation | INTC | 14.15% |
| BlackRock Inc | BLK | 11.24% |
| Total SA | TOT | 7.90% |
| T Rowe Price Group Inc | TROW | 7.81% |
| Comcast Corporation | CMCSA | 6.73% |
| Caterpillar Inc | CAT | 5.99% |
| FedEx Corp | FDX | 5.57% |
| Exxon Mobil Corp | XOM | 5.01% |
| Chevron Corporation | CVX | 4.89% |
| Goldman Sachs Group Inc | GS | 4.88% |
| Motorola Solutions Inc | MSI | 4.14% |
| BB and T Corporation | BBT | 3.45% |
| Simon Property Group Inc | SPG | 2.92% |
| Cognizant Technology Solutions Corp | CTSH | 2.57% |
| Verizon Communications Inc | VZ | 2.51% |
| Deere and Co | DE | 2.46% |
| JP Morgan Chase and Co | JPM | 2.45% |
| Pfizer Inc | PFE | 2.40% |
| Avalonbay Communities Inc | AVB | 1.50% |
| Discover Financial Services | DFS | 1.10% |
| Air Products and Chemicals Inc | APD | 0.35% |
| 100.00% |
Market Efficiency in an Irrational World
Traditionally, economic theory takes the view that investors are rational; but several behavioral biases such as Overconfidence have been studied recently. In particular, this overconfidence has been found to indirectly cause investors to underweight or overweight new information (caused by the "cognitive dissonance", "attribution bias", and "conservatism bias"), contributing to the momentum effect. The authors propose that growth stocks (i.e., low book-to-market ratios) should exhibit higher momentum due to their value being less predictable (e.g., most of their assets are intangible) than more stable stocks.
The authors study the July 1963 - December 1997 period in the United States equity markets and find that stocks with high Book-to-Market characteristics (i.e., value stocks) and high TTM returns (i.e., momentum stocks) significantly outperformed their growth and low momentum counterparts, as well as the market portfolio. The value/momentum portfolio outperformed the market portfolio by almost 0.60% per month.
The authors introduce "adaptive efficiency", which relaxes the efficient market hypothesis by adding behavioral theory. They suggest that the "irrational investors" may tilt their portfolios to anomalies, while the "rational investors" assume that anomalies are corrected by rational investors and will not try to exploit the anomalies; and this will cause the anomalies (such as momentum) to persist. The authors prove this to be the case over the 1974 - 1997 period in the United States equity markets.
DANIEL, K., & TITMAN, S. (1999). Market Efficiency in an Irrational World. Financial Analysts Journal, 55.
The authors study the July 1963 - December 1997 period in the United States equity markets and find that stocks with high Book-to-Market characteristics (i.e., value stocks) and high TTM returns (i.e., momentum stocks) significantly outperformed their growth and low momentum counterparts, as well as the market portfolio. The value/momentum portfolio outperformed the market portfolio by almost 0.60% per month.
The authors introduce "adaptive efficiency", which relaxes the efficient market hypothesis by adding behavioral theory. They suggest that the "irrational investors" may tilt their portfolios to anomalies, while the "rational investors" assume that anomalies are corrected by rational investors and will not try to exploit the anomalies; and this will cause the anomalies (such as momentum) to persist. The authors prove this to be the case over the 1974 - 1997 period in the United States equity markets.
DANIEL, K., & TITMAN, S. (1999). Market Efficiency in an Irrational World. Financial Analysts Journal, 55.
Thursday, October 11, 2018
Cross-sectional and Time-series Determinants of Momentum Returns
Previous studies have found that previous winners over a 3-12 month period will show 1%/month profits over the next 12 months. This has been empirically confirmed across several markets and time periods, even back to the 1920s.
Several reasons for this phenomenon to occur have been proposed: under-reaction to information; delayed over-reaction to information; or an undetected risk. In particular Conrad and Karl in a previous study find it to be due to cross-sectional dispersion in unconditional expected returns.
The authors find Conrad and Karl's study to be flawed due to a small sample bias, and they prove that their conclusion explains very little, if any, of the momentum profits. In fact, they say "virtually none of the momentum profits can be attributed to compensation for risk."
Jegadeesh, N., & Titman, S. (2002). Cross-Sectional and Time-Series Determinants of Momentum Returns. Review of Financial Studies, 15(1), 143–157.
Several reasons for this phenomenon to occur have been proposed: under-reaction to information; delayed over-reaction to information; or an undetected risk. In particular Conrad and Karl in a previous study find it to be due to cross-sectional dispersion in unconditional expected returns.
The authors find Conrad and Karl's study to be flawed due to a small sample bias, and they prove that their conclusion explains very little, if any, of the momentum profits. In fact, they say "virtually none of the momentum profits can be attributed to compensation for risk."
Jegadeesh, N., & Titman, S. (2002). Cross-Sectional and Time-Series Determinants of Momentum Returns. Review of Financial Studies, 15(1), 143–157.
Wednesday, October 10, 2018
The Profitability of Momentum Strategies
The authors study the effects of price and earnings momentum over the period 1973-1993 in the United States equity market. They find that winners over the past 6 months significantly outperform losers over the next 6-12 months.
Drilling in, they find that price momentum produces better returns for longer holding periods than earnings momentum. They contribute this to the theory that earnings is more of a short-term measure; whereas price changes could be due to very long-term changes. They even found these things to be true for large-cap stocks, which would be expected to not exhibit as much momentum capture due to their better and more public information than that of small-caps.
They contribute this effect to several possibilities: the market does not fully respond to new information, due to investors' conservatism bias (where they are reluctant to change prior opinions); or maybe by analysts being slow to revise estimates. They note that the momentum effect is not caused by the trades of trend chasers, because there is no subsequent reversal to bring the stock back to equilibrium (even out to the 3rd year).
Chan, L. K. C., Jegadeesh, N., & Lakonishok, J. (1999). The Profitability of Momentum Strategies. Financial Analysts Journal, 55(6), 80.
Drilling in, they find that price momentum produces better returns for longer holding periods than earnings momentum. They contribute this to the theory that earnings is more of a short-term measure; whereas price changes could be due to very long-term changes. They even found these things to be true for large-cap stocks, which would be expected to not exhibit as much momentum capture due to their better and more public information than that of small-caps.
They contribute this effect to several possibilities: the market does not fully respond to new information, due to investors' conservatism bias (where they are reluctant to change prior opinions); or maybe by analysts being slow to revise estimates. They note that the momentum effect is not caused by the trades of trend chasers, because there is no subsequent reversal to bring the stock back to equilibrium (even out to the 3rd year).
Chan, L. K. C., Jegadeesh, N., & Lakonishok, J. (1999). The Profitability of Momentum Strategies. Financial Analysts Journal, 55(6), 80.
Tuesday, October 9, 2018
Value versus growth stocks and earnings growth in style investing strategies in Euro-markets
The authors study the Euro-zone during the period 1988-2003. They first corroborate prior studies that show value strategies outperform growth strategies. They then go on to define undervalued/(overvalued) value/(growth) stocks as those with high(low) EPS growth and high(low) E/P or B/P ratios, respectively. They further show that undervalued value stocks outperform overvalued growth stocks over the period.
The authors further analyze the effect of expected EPS changes on value and growth strategies, following a study by Ahmed and Nanda in 2001. The authors corroborate that study by finding that stocks with positive earnings momentum outperform those with negative earnings momentum, especially when combined with the undervalued value stocks.
Chahine, S. (2009). Value versus growth stocks and earnings growth in style investing strategies in Euro-markets. Journal of Asset Management, 9(5), 347–358.
The authors further analyze the effect of expected EPS changes on value and growth strategies, following a study by Ahmed and Nanda in 2001. The authors corroborate that study by finding that stocks with positive earnings momentum outperform those with negative earnings momentum, especially when combined with the undervalued value stocks.
Chahine, S. (2009). Value versus growth stocks and earnings growth in style investing strategies in Euro-markets. Journal of Asset Management, 9(5), 347–358.
Monday, October 8, 2018
A tale of two states: asymmetries in the UK small, value and momentum premiums
The authors study the asymmetries in the value, size, and momentum factor premiums across expansionary and recessionary regimes in the UK over the period January 1982 - June 2014. They find strong evidence for asymmetries across all three factors in the different regimes; the size premium showed the strongest asymmetry (even turning negative in the recessionary regime), and the momentum factor showed the least asymmetry.
The authors also corroborate US studies by finding that macroeconomic factors (such as GDP Growth, credit, inflation, etc.) are significant determinants of value/size/momentum factor premiums, especially in market downturns. The authors also developed a Markov switching model and back-tested a portfolio, beating a buy and hold strategy across 8 size/value/momentum portfolios.
Sarwar, G., Mateus, C., & Todorovic, N. (2017). A tale of two states: asymmetries in the UK small, value and momentum premiums. Applied Economics, 49(5), 456–476.
The authors also corroborate US studies by finding that macroeconomic factors (such as GDP Growth, credit, inflation, etc.) are significant determinants of value/size/momentum factor premiums, especially in market downturns. The authors also developed a Markov switching model and back-tested a portfolio, beating a buy and hold strategy across 8 size/value/momentum portfolios.
Sarwar, G., Mateus, C., & Todorovic, N. (2017). A tale of two states: asymmetries in the UK small, value and momentum premiums. Applied Economics, 49(5), 456–476.
Saturday, October 6, 2018
Portfolio - October 6, 2018
| Company Name | Ticker | Weight |
| Intel Corporation | INTC | 18.16% |
| BlackRock Inc | BLK | 11.02% |
| CVS Caremark Corporation | CVS | 10.47% |
| Comcast Corporation | CMCSA | 10.05% |
| Total SA | TOT | 10.03% |
| T Rowe Price Group Inc | TROW | 6.82% |
| Goldman Sachs Group Inc | GS | 5.10% |
| FedEx Corp | FDX | 4.82% |
| Simon Property Group Inc | SPG | 4.36% |
| Chevron Corporation | CVX | 4.05% |
| Motorola Solutions Inc | MSI | 3.94% |
| BB and T Corporation | BBT | 2.92% |
| Cognizant Technology Solutions Corp | CTSH | 2.27% |
| PACCAR Inc | PCAR | 2.11% |
| Texas Instruments Incorporated | TXN | 1.25% |
| Deere and Co | DE | 1.09% |
| Avalonbay Communities Inc | AVB | 0.99% |
| Discover Financial Services | DFS | 0.55% |
| 100.00% |
Thursday, October 4, 2018
Size, Value, Momentum in Indian Equities
The authors study the Indian equity markets during the period January 1994 - March 2017. In doing so they seek to learn the performance of the value (HML), size (SMB), and momentum indicators (12-1) in that market over that time period.
They used the Fama-French methodology to form portfolios and determined that value significantly outperformed growth and winners significantly outperformed losers; however, small cap significantly underperformed large cap, which is contrary to the Fama French and other studies.
The HML factor returned 9.08% with an annualized volatility of 16.33%, and a max draw-down of 53%. The size factor returned 0.36% with an annualized volatility of 14.51% and a max drawdown of 74% in 1995 that never recovered. The momentum factor returned 17.3% with an annualized volatility of 17.06%, and a max draw-down of 49%.
The authors further mean-variance optimized the factors with market factors and a risk-free rate; interestingly, the portfolios at moderate/high risk levels primarily held momentum and value factors, while lower risk portfolios primarily held the risk-free security. The market factor was a small weight in all portfolios.
Agarwalla, S. K., Jacob, J., & Varma, J. R. (2017). Size, Value, and Momentum in Indian Equities. Vikalpa: The Journal for Decision Makers, 42(4), 211–219.
They used the Fama-French methodology to form portfolios and determined that value significantly outperformed growth and winners significantly outperformed losers; however, small cap significantly underperformed large cap, which is contrary to the Fama French and other studies.
The HML factor returned 9.08% with an annualized volatility of 16.33%, and a max draw-down of 53%. The size factor returned 0.36% with an annualized volatility of 14.51% and a max drawdown of 74% in 1995 that never recovered. The momentum factor returned 17.3% with an annualized volatility of 17.06%, and a max draw-down of 49%.
The authors further mean-variance optimized the factors with market factors and a risk-free rate; interestingly, the portfolios at moderate/high risk levels primarily held momentum and value factors, while lower risk portfolios primarily held the risk-free security. The market factor was a small weight in all portfolios.
Agarwalla, S. K., Jacob, J., & Varma, J. R. (2017). Size, Value, and Momentum in Indian Equities. Vikalpa: The Journal for Decision Makers, 42(4), 211–219.
Wednesday, October 3, 2018
Enhancement of value portfolio performance using momentum and the long-short strategy: The Finnish evidence
The authors continue the work of Bird and Casavecchia to study the performance of portfolios composed of both value and momentum stocks. They use the Finnish stock market returns of 1993 to 2008 and form value/momentum portfolios. They find that portfolios with a composite of valuation indicators (i.e., D/P, B/P, and EBITDA/EV) and top quintile 6-month price momentum outperformed the market by nearly 5 percentage points annually, and the volatility was decreased by 0.86 percentage points.
They went on to study the effect of using this portfolio in a 130/30 long/short strategy, thereby shorting 30% of the "glamour losers" and using those proceeds to go long 30% "value winners". This strategy produced returns that doubled the market returns and lowered volatility by 3% points.
Leivo, T. H., & Pätäri, E. J. (2011). Enhancement of value portfolio performance using momentum and the long-short strategy: The Finnish evidence. Journal of Asset Management, 11(6), 401–416.
They went on to study the effect of using this portfolio in a 130/30 long/short strategy, thereby shorting 30% of the "glamour losers" and using those proceeds to go long 30% "value winners". This strategy produced returns that doubled the market returns and lowered volatility by 3% points.
Leivo, T. H., & Pätäri, E. J. (2011). Enhancement of value portfolio performance using momentum and the long-short strategy: The Finnish evidence. Journal of Asset Management, 11(6), 401–416.
Tuesday, October 2, 2018
The performance of value and momentum investment portfolios: Recent experience in the major European markets
The authors use the equity returns of major European markets over the period 1990 - 2002 to study various value and momentum strategies.
For value, they found that the use of book-to-market and sales-to-price criterion perform better than dividend yield and earnings yield. Even when adjusting for size and country bias, "cheaper" stocks significantly outperform the "expensive" stocks.
For momentum, they looked at price momentum (on a 6-month and 12-month criteria) and earnings momentum (agreement measures and forecast revisions). Even when adjusting for size and country bias, "past winners" significantly outperform "past losers". The past winners are typically "growth" and large-cap biased, and they perform best at 6-12 month holding periods.
Bird, R., & Whitaker, J. (2003). The performance of value and momentum investment portfolios: Recent experience in the major European markets. Journal of Asset Management, 4(4), 221–246.
For value, they found that the use of book-to-market and sales-to-price criterion perform better than dividend yield and earnings yield. Even when adjusting for size and country bias, "cheaper" stocks significantly outperform the "expensive" stocks.
For momentum, they looked at price momentum (on a 6-month and 12-month criteria) and earnings momentum (agreement measures and forecast revisions). Even when adjusting for size and country bias, "past winners" significantly outperform "past losers". The past winners are typically "growth" and large-cap biased, and they perform best at 6-12 month holding periods.
Bird, R., & Whitaker, J. (2003). The performance of value and momentum investment portfolios: Recent experience in the major European markets. Journal of Asset Management, 4(4), 221–246.
Monday, October 1, 2018
The performance of value and momentum investment portfolios
The authors build portfolios of stocks that are characterized by both value and momentum factors and determine they outperform. The thought is that value and momentum factors are opposing (i.e., momentum is typically associated with growth rather than value stocks). By selecting stocks with both attributes, they form portfolios with components that are less correlated; and in addition, they buy cheap stocks when they are on the rise, which may be a more appropriate time to buy them than when they are falling.
Bird, R., & Whitaker, J. (2004). The performance of value and momentum investment portfolios: Recent experience in the major European markets Part 2. Journal of Asset Management, 5(3), 157–175.
Bird, R., & Whitaker, J. (2004). The performance of value and momentum investment portfolios: Recent experience in the major European markets Part 2. Journal of Asset Management, 5(3), 157–175.
Subscribe to:
Posts (Atom)