Exploring Consumer Sentiment in Predictive Modeling

598 views

Published on

Exploring the utility of Bloomberg Consumer Comfort Index in predicting stock market trends

Published in: Economy & Finance, Business
0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total views
598
On SlideShare
0
From Embeds
0
Number of Embeds
232
Actions
Shares
0
Downloads
0
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide

Exploring Consumer Sentiment in Predictive Modeling

  1. 1. Annual conference of the American Association for Public Opinion Research Anaheim, California – May 16, 2014 Pavle Gegaj, Gary Langer and Julie E. Phelan Contact: glanger@langerresearch.com
  2. 2.  Continuous weekly tracking of consumer sentiment since December 1985  N=250/week summed in a four-week rolling avg.  Recently rebased as avg. pos. ratings on nat’l economy, buying climate, personal finances; range 0-100  Followed in financial markets and by economists and econometricians  RDD telephone survey by SSRS; 50% cell, 50- state, English/Spanish, Census weights
  3. 3. • The Bloomberg Consumer Comfort Index: Concurrent and Predictive Validity (Phelan/Langer) ◦ Presented at AAPOR 2011 ◦ Evaluated face validity, internal consistency (Cronbach’s α = .85), convergent validity, known- groups validity, concurrent validity and predictive validity of the CCI in comparison with other indicators ◦ See also “Adding Cell Phone Sampling to a Long- Term Trend Study: The Bloomberg Consumer Comfort Index” (Phelan/Langer AAPOR 2013)
  4. 4.  Does the CCI add explanatory power to the traditional model predicting personal expenditures?  Does the CCI anticipate stock market trends?
  5. 5. CCI Conference Board Michigan CCI Economy CCI Finances CCI Buying Climate Overall index CCI subindices Personal expenditures overall .1941 .2456 .2493 .1747 .1795 .1614 Motor vehicles .0809 .0353 .0155 .0920 .0063 .0367 Goods, excl. motor vehicles .0865 .1143 .0959 .0671 .0915 .0778 Services .1918 .3087 .3330 .1826 .1856 .1716 Durables, excl. motor vehicles .1504 .1463 .1375 .1090 .0910 .1727 Lagged consumer sentiment explains substantial variance in personal expenditures
  6. 6. CCI Conference Board Michigan CCI Economy CCI Finances CCI Buying Climate Overall index CCI subindices Personal expenditures overall .1937 .2872*** .2736** .1774 .2291*** .1540 Motor vehicles .0713 .0512 .0397* .0877 .0063 .0331 Goods, excl. motor vehicles .0801 .1334* .0987 .0760 .1101* .0733 Services .2010 .3261* .3401 .1904 .2348*** .1696 Durables, excl. motor vehicles *** = p < .01 ** = p < .05, * p < .1 .2625*** .2622*** .2496*** .2163*** .2278*** .2355*** Adding contemporary values improves the variance explained
  7. 7.  Personal income  S&P 500  Three-month T-bill rate  Unemployment  Lagged personal expenditures …to which we add consumer sentiment
  8. 8. CCI Conference Board Index Michigan Index Economy Index Finance Index Buying Index Increase of adjusted R2 Overall index CCI subindicies Personal expenditures overall -.0006 .0067 .0170 -.0041 -.0052 .0049 Motor vehicles .0218* .0368* .0059 .0118 .0383* .0457** Goods, excl. motor vehicles .0559** -.0032 .0153* .0195* -.0021 .0702*** Services -.0051 -.0001 .0028 -.0066 -.0069 -.0061 Durables, excl. motor vehicles .0733*** .0151* .0443* .0342* .0455** .0983*** Adding consumer sentiment to the traditional model, particularly CCI, explains additional variance in predicting personal expenditures
  9. 9.  Brusuelas/Lawrie in Bloomberg Brief, 5/6/14  On a 9-month lead, separate CCI measure of economic expectations correlates with personal expenditures at .65  Conclusion: “Bloomberg index helps forecast personal spending”
  10. 10.  Simple model: ◦ Hypothetically buy one stock unit on any rise in each week’s CCI ◦ Sell all units on any fall in the CCI ◦ Hold on no change  Tested using CCI and stock market data from mid-2004 to 2012
  11. 11. June 2004 - January 2012 Asset CCI Buy and hold Dow Jones 134.75% 22.51% Microsoft 144.07% 32.53% Exxon 243.92% 126.69% Walmart -141.46% 29.64% General Electric 170.79% -24.11% Coca-Cola 112.47% 65.03% Total S.A. 248.66% 57.13% McDonald’s -153.86% 356.63% PepsiCo. 90.45% 44.08%
  12. 12. Buy and hold S&P 500 vs.:  a) CCI long/flat: buy on increase in CCI, sell all on decrease  b) CCI long/short: buy on increase in CCI, short on decrease  Tested for period 2000-2014
  13. 13.  Buy/hold: 28% return, average annual 1.7%, Sharpe ratio (risk-adjusted return) -0.01.  CCI long/flat: 104% return, annual average 5.1%, Sharpe ratio 0.28.  CCI long/short: 133% return, annual average 6.0%, Sharpe ratio 0.22.
  14. 14. The Bloomberg Consumer Comfort Index is produced by Langer Research Associates LLC (Langer).CONSUMER COMFORT INDEX ™ is a trademark or service mark of Langer. BLOOMBERG ® is a trademark or service mark of Bloomberg Finance L.P. Bloomberg Finance L.P. and its affiliates (collectively, "Bloomberg") or Bloomberg's licensors own all proprietary right in such mark. Neither Langer nor Bloomberg guarantee the timeliness, accuracy or completeness of any data or information relating to the Bloomberg Consumer Comfort Index. Langer and Bloomberg makes no warranty, express or implied, as to the Bloomberg Consumer Comfort Index or any data or values relating thereto or results to be obtained therefrom, and expressly disclaims all warranties of merchantability and fitness for a particular purpose with respect thereto. To the maximum extent allowed by law, Langer, Bloomberg, their licensors, and their respective employees, contractors, agents, suppliers and vendors shall have no liability or responsibility whatsoever for any injury or damages - whether direct, indirect, consequential, incidental, punitive or otherwise - arising in connection with the Bloomberg Consumer Comfort Index or any data or values relating thereto - whether arising from their negligence or otherwise. Nothing in the Bloomberg Consumer Comfort Index shall constitute or be construed as an offering of financial instruments or as investment advice or investment recommendations (i.e., recommendations as to whether or not to “buy,” “sell,” “hold” or to enter or not to enter into any other transaction involving any specific interest or interests) by Langer, Bloomberg or its affiliates or a recommendation as to an investment or other strategy by Langer, Bloomberg or its affiliates.
  15. 15.  Explore the extent to which the CCI improves existing models predicting economic indicators other than personal expenditures  Extend stock market modeling to: ◦ Use other time periods, moving averages and more complex trading approaches ◦ Use additional indices, specific sectors and individual companies and bond yields ◦ Test CCI subindices as well as the full index ◦ Further examine positive vs. negative results
  16. 16. Annual conference of the American Association for Public Opinion Research Anaheim, California – May 16, 2014 Pavle Gegaj, Gary Langer and Julie E. Phelan Contact: glanger@langerresearch.com Thank you!

×