The+Golden+Question+Answered

//[|Gold vs. S&P Motion Chart]//
= =

=The Golden Question Answered= By: Ali Alyabis, Andrew Clark, & David Lee

= Final Stream Snap Shot = = =

__Introduction:__
Conventional wisdom holds that the stock market and gold prices are inversely related. When expectations of the stock market are low, people put their money in gold. When the market is expected to rise, gold prices typically fall. This study will examine whether or not macroeconomic conditions and indices, have any bearing on this relationship.

Specifically, the study will examine the importance of American socioeconomic conditions on the relative prices of Gold to stocks. Accordingly, we will examine commodity and energy prices, inflation and unemployment rates, political power and the propensity of companies to go public. While seemingly unrelated, we hope to find an unknown relationship in the data.

=__Actual Steps for Data Collection and Project__=  -Data Mining followed CRISP guidelines

-Gather data -Clean data -Build models -Interpret results -Build and test model

-Find additional data -Clean data -Build models -Interpret results -Build and test final model -Explore literature further and talk to professionals



__Further Data Collection and Analysis Beyond Project Proposal__
We were unable to find any relationship in our original data. Accordingly, we expanded our dataset to include the below variables. Originally we were looking for economic indicators of S & P 500 performance including market basket items, fuel prices, IPOs, population, political indicators, and recession information. Consequently after further review it came to our attention there would be pretty predictors than the variables originally researched. The following variable names, description of the variable and variable type are included in the table below:


 * ~ **Variable ** ||~ **Description ** ||~ **Type ** ||
 * = Date ||= Numeric Date ||= Date ||
 * = Year ||= Year between 1970 and 2009 ||= Range ||
 * = Month ||= January to December (1-12) ||= Set ||
 * = S&P_CLOSE ||= S&P Index at Month End (in US Dollars) ||= Range ||
 * = Gold Prices ||= Price of Gold per Ounce (in US Dollars) ||= Range ||
 * = AA ||= Stock Price of Alcoa at Month End (in US Dollars) ||= Range ||
 * = MCD ||= Stock Price of McDonalds at Month End (in US Dollars) ||= Range ||
 * = JNJ ||= Stock Price of Johnson and Johnson at Month End (in US Dollars) ||= Range ||
 * = Unemployment ||= Unemployed Percentage of US labor force ||= Range ||
 * = Inflation ||= US Inflation Rate (Percentage Points) ||= Range ||
 * = Misery Index ||= Inflation Rate plus Unemployment Rate (Percentage Points) ||= Range ||
 * = 1 Month CD ||= Average Rate of One Month CD ||= Range ||
 * = Expected VS REAL ||= If percent change of 1 Month CD, is greater than percent change in Inflation then Expected. If not, Real. ||= Binary ||
 * = Consumer Sentiment ||= University of Michigan Consumer sentiment Index (Percentage Points) ||= Range ||
 * = GOLD OR S&P ||= If percent change of in Gold Price is greater than percent change in S&P_Close, then Gold. If not, S&P. ||= Binary ||


There was little predictive merit in any of the models. The C&R Tree was the most accurate at 66.52%, meaning it correctly identified opportunities to invest in Gold or the S&P 500 less than two thirds of the time.

The C&R Tree, gave us some indication of an investment strategy for today (May 2009), given the high national unemployment rate. With the unemployment rate over 7.45% and the 1-Month CD below 5.05%, history would suggest now is the time to buy Gold. That is to say, in twelve of the past fifteen months with comparable unemployment and 1-Month CD rates, Gold has proven more profitable. However, before rushing to gold, it should be noted that these are unprecedented times. The last two months (April and May of 2009), were unlike any other in history and the American Economy retracted to the point of negative inflation. In both months it proved more profitable to invest in the S&P. Thus there is evidence supporting an immediate

__ Other Findings: __
While mining for Gold growth, a few additional interesting nuggets emerged. Specifically, when inflation is very low (i.e. 0.05 or less) it is always better to hold gold until the inflation rate goes below zero. In March and April of this year, this situation has been the case. And contrary to popular wisdom, it would have been better to hold stocks in these months rather than sell.

The following cluster information is a finding from a Two-Step Hierarchical Cluster Analysis. We found natural time clusters. The information is displayed below along with our cluster names.







Time Cluster Names

 * = ** Cluster ** ||= ** Name ** ||= ** Time Period ** ||= <span style="display: block; background-color: rgb(128, 128, 128); text-align: center;">** Major Event ** ||
 * = 1 ||= “Panned Out” ||= 1978 - 1982 ||= Stagflation ||
 * = 2 ||= “Prospectors” ||= 1983 - 1988 ||= Ali Born ||
 * = 5 ||= “The Golden Era” ||= 1989 - 1997 ||= Berlin Wall Falls ||
 * = 4 ||= “Mine Kamph” ||= 1998 - 2001 ||= .dot Com Bubble and Y2K ||
 * = 3 ||= “Gold Digs, to Brokety Broke” ||= 2002 - 2008 ||= Crisis from Authority ||

__//Gold September://__
Through Apriori association, we discovered that consumer sentiment most often dropped in the month of September. Interestingly, but by no means necessarily related, we also found that over the past decade it has proven more profitable to hold Gold in the month. Please see the chart below:



=__Implications for Investors:__=

As discussed above, we had difficulty determining strong relationships or any broader macroeconomic indication from our models. Significantl more work research would need to be done to come up with an sound, actionable investment strategy.

However, if we were forced to invest $1,000 at the start of June of 2009, Our strategy would be as follows:

//Central Assumption//: As of May 2009, the US economy is weak. Barring some major economic event or major change in policy, it is reasonable to expect inflation and economic growth to be very low, even negative. For the sake of discussion we assume a range of -2% - 2%.


 * If we expect inflation to be negative (less than 0%), we would invest in S&P.

If inflation was expected to be positive (between 0 and 2%), then we'd buy Gold.**

__**Annual Inflation Rate:**__
=__Ask the Experts__= Members of the team consulted Professor Dothan about our results. He suggested that conventional investment wisdom is; gold rises when inflation beats market expectation of inflation. While our results were not strong enough to make strong predictions or relationships to this suggestion, we do not find data that contradicts this conventional wisdom either.

=__Reporting on Errors__= Our variables showed very little relationship between one another. While the data we collected was very comprehensive, it was still insufficient, specifically we would have adjusted our variable for inflation expectations into something that was more along the lines of what we wanted. We used a 1-month CD rate when there is probably a better treasury bill or treasury security that is more predictive of inflation.

=__Further Research__= Reporting on investment strategies is a difficult task. Researchers and mathematicians have looked for patterns in the stock market for years. The nature of the stock market is very sensitive to inflation and economic expectations. Clearly, we need variables that are more predictive of investment and market behaviors in order to give a solid recommendation to investors. But this is extremely difficult. But mining through massive amounts of data often times very small patterns emerge. This is indeed the case, but as mentioned before, no results yeilded strong investment recommendations.

<span style="font-size: 14pt; color: rgb(120, 172, 53); font-family: Georgia;"> References:

Article - [] <span style="font-size: 14pt; color: rgb(120, 172, 53); font-family: Georgia;"> Article - [] <span style="font-size: 14pt; color: rgb(120, 172, 53); font-family: Georgia;"> “Consumer Expenditure Surveys.” US Bureau of Labor Statistics (BLS) Homepage. Accessed 2 April 2009. [] <span style="font-size: 14pt; color: rgb(120, 172, 53); font-family: Georgia;"> “Economic Data--(FRED)®” Economic Research: Federal Reserve Bank of St. Louis. Accessed 2 April 2009. [] <span style="font-size: 14pt; color: rgb(120, 172, 53); font-family: Georgia;"> “Historic Data for S&P 500 Stocks” Accessed 9 April 2009. [] <span style="font-size: 14pt; color: rgb(120, 172, 53); font-family: Georgia;"> Ritter, John R. “Founding dates for 8,823 IPOs from 1975-2008.” IPO Data. October 2008. Accessed 9 April 2009. []

Daly, Ryan M,"Tactical Asset Allocation to Gold" (May 2005). Accessed 12 May 2009 Available at [|http://ssrn.com/abstract=783187]

McMahon, Timothy " Is Inflation Rising or Falling? " May 12 2009. Accessed 12 May 2009. []

Adrian, Ash " Dow Gold Ratio Hits 80 Year Extremes, Time to Buy?" Accessed 12 May 2009. []