How the Prominent Survey of Financial Forecasts data is transformed from low-frequency survey data into weekly high-frequency insights.
Our basic data comes from a well-known monthly publication which reports the survey results of forecasts by over 40 financial institutions and other experts of a wide range of interest rates and macroeconomic variables. These forecasts are quarterly averages and go out six quarters into the future. Because this publication reports the survey results almost 3 weeks after the survey was undertaken, events may have eclipsed the value of the forecasts.
Using a statistical technique pioneered by Professor Ghysels and other econometricians, we can turn low-frequency data (monthly) into high-frequency data (weekly). Practically, this means we can now issue quarterly forecasts on a weekly rather than monthly basis.
By tracking these forecasts back-to-front, we can see how views changed over time. Presumably, 6-quarter-out forecasts will be very diffuse with a wide variety of opinions possible. As the forecast horizon shortens from 6 quarters out to 5, 4, … until the current quarter, the number of viable alternative views will coalesce until there is only one major view or forecast.
In statistical terms, the diffuse 6-quarter-out forecast will go from a multimodal histogram of forecasts to a unimodal histogram. The knowledge of when opinion is starting to coalesce should be important information for many finance professionals and traders — it gives them information of when to buy and sell interest rate securities.
One of the key financial variables that are the focus of financial professionals is the Federal Funds Rate — the short-term interest rate targeted by the Federal Open Market Committee (FOMC). Our analysis examines the views of over 40 market forecasters as to what the Federal Funds Rate is justified by the current state of the economy.
In a 1993 paper, John Taylor introduced an equation that prescribes a value for the Federal Funds Rate based on the values of inflation and economic slack (output gap or unemployment gap). Since then, alternative versions of Taylor's original equation have been used and called "simple monetary policy rules" or "modified Taylor rules." One application of our weekly quarterly forecast model results in forecasts of Federal Funds Rate prescriptions based on a generalization of Taylor's original formula.
In addition, our analysis calculates the probabilities of an increase in the Federal Funds Rate based on a paper by Prof. Melick on Recovering Fed Funds Target Rate Probabilities from Fed Funds futures options. Opinions will vary widely 6 quarters out, only to ultimately converge to one dominant opinion. Knowing when this is about to happen is important information for many finance professionals and traders.
Every capitalist economy undergoes business cycles with expansion and contraction phases. It is critical to anticipate turning points from expansion to recession and back to expansion. We offer three independent methods to calculate the likelihood of a recession.
Signals the start of a recession when the three-month moving average of the national unemployment rate (U3) rises by 0.50 percentage points or more relative to the minimum of the three-month averages from the previous 12 months.
Neftci proposed using sequential analysis to calculate the probability of a cyclical turning point. Sequential analysis evaluates data as it is collected — a form of Bayesian analysis that incorporates and updates the model with new evidence. Neftci's method is based on the empirical claim that the onset of a recession is marked by a pronounced decline in aggregate economic activity.
An inverted yield curve (negative 10Y–3M spread) has preceded every US recession historically, typically by 12–18 months. Our model converts the spread into an explicit 12-month recession probability.
Stocks are basically valued based on discounted earnings. The discount factor is an interest rate based on a variety of years to maturity. Given the mix of interest rates and macroeconomic variables in our weekly pseudo surveys, our data can be used to explore a wide range of stock prices. In this application, we have focused on the S&P 600 index of stock prices as well as its 11 GICS sector sub-components including Health Care, Financials, and Industrials.
As with the survey data, by tracking these forecasts back-to-front, we can see how views changed over time. Six-quarter-out forecasts will be very diffuse, reflecting a wide variety of opinions. As the forecast horizon shortens, the number of viable alternative views will coalesce until there is only one major view or forecast. Knowing when this is about to happen is important information for many finance professionals and traders — it gives them valuable information on when to buy and sell.
The financial news frequently compares the latest business data to market expectations. The number of experts we survey weekly allows us to provide pseudo forecasts of their thinking for a broader set of economic indicators beyond the core survey variables.
In addition to possibly having more respondents and providing their views on a 6-quarter-out to current-quarter horizon, as with the previously described applications, we can see how these views have changed over time. Six-quarter-out forecasts will again be very diffuse. As the forecast horizon shortens, the number of viable alternative views will coalesce until there is only one major view. Knowing when this is about to happen gives finance professionals and traders valuable trading information.
Creating the optimal bond portfolio involves two steps: (a) constructing the efficient frontier and (b) choosing the portfolio that lies on that frontier and satisfies the investor's objectives and constraints. The efficient frontier is the set of portfolios that offer the highest possible return for a given level of risk, or the lowest possible risk for a given level of return.
We use the wide range of forecasts over a variety of interest rates and maturities as our measure of risk. From the positive slope of the Efficient Frontier, investors who choose higher returns must be willing to absorb more risk.
Defining constraints includes specifying the target return and time horizon or holding period desired. Holding periods depend on liquidity needs, tax considerations, and other relevant factors. With 6-quarter-out forecasts, we offer the user a wide choice of holding periods stretching well into the future.
Because our choices of financial assets include government bonds, corporate bonds, municipal bonds, and more, our portfolios allow for a rich choice of holdings. Finally, because we track these portfolios over time, we allow the user to rebalance — adjusting portfolio weights to maintain the desired risk-return profile as new Prominent Survey data arrives each week.