Last week David Kostin, a well respected Goldman Sachs' chief U.S. equity strategist published his forecast for 2012-2013. He thinks the S&P 500 (SPY) will end 2012 at 1,250, then rise in 2013. Here's how he sees the S&P moving through the end of next year:
Market forecasting is dangerous. You put something in print, and pray that a year later you could say, "See, I told you so". Meanwhile wars, droughts, tsunamis and banker's follies all conspire to make you subject to ridicule.
We present in this article our forecast for S&P 500 index from October 12. We don't know what Goldman Sachs (GS) forecast is based on. Thus our forecast is completely independent of theirs.
How does our prediction model make forecasts?
The prediction system is based on the realization that the market as a whole, and individual components, stocks, bonds, commodities and currencies, all move in waves. The money is constantly moving between thousands of investment venues looking for better return. Thus all the markets are interconnected and what happens in one market affects the rest.
The current value of S&P 500 index is a precise function of the value of each of its 500 components. However, its future value depends indirectly on much more. The crisis in Europe, the presidential elections, and other items in the news feed affect the expectations of the future value of S&P 500. Stock market forecast algorithms are able to analyze these inputs, arrange them in the order of importance, (principal component analysis, PCA), and build from them a number of mutually competing models. Each model projects the future trajectory of the given market in the multidimensional space of other markets. The system outputs the predicted trend as a number, positive or negative, along with the wave chart that predicts how the waves will overlap the trend. The next step is to combine these models in a statistically coherent super-model.
It's important to understand that forecasts can't be static. Each day brings new information into the market, and it must be reflected in the forecast.
Moreover, each forecasting model is a set of rules. But the rules, the forces of the market are not static either. What was important yesterday may not be today. Thus the model has to be adapted to reflect the new reality, but must still follow the general historical rules. The rules are derived from historical data which is constantly regurgitated by high-power computer systems. Many different scenarios are proposed and tested. The "fittest" models are combined, mutated and re-tested through a process inspired by nature. This highly repetitive process is aptly named "Genetic Algorithms". The competing models are rated by a complex scoring system according to the ability to predict the future. The best model needs to perform not only on the recent data, but also on any given subset of historical data.
Another important indicator is the predictabilty: just like the market rises and falls in waves, so does the predictability. It moves in waves. And the waves are not synchronous. The focus of market attention constantly shifts between different venues, be it gold, stocks, oil, bonds. Some indexes become more predictable, while the others retreat to randomness. Global trends and events affect the forecast-ability of various indexes. As of today S&P 500 index is not the most forecast-able one. There are stocks and indices with higher predictability.
We believe that the S&P 500 is no longer a good forward indicator of market health. It includes the old heavy industry companies that will lag behind the more nimble hi-tech. In today's economy NASDAQ is a better indicator, and we see it outperform the S&P 500 in 2013. The broad market index Russell 2000 (RUT) will also outperform the S&P 500.
Our S&P 500 forecast:
So here is our S&P 500 forecast for the end of 2012 based on I Know First predictive algorithm. It calls for a brief uptrend (low predictability), then a correction downward towards the end of the year (higher predictability). The next year (not on chart) will be restricted to range trading with somewhat downward bias:
(click to enlarge)
So basically, our forecast agrees with the Goldman's, at least for the near term.
More forecasts farther into 2013, in broad brush strokes.
We might be off in some of these forecasts. We present them here because we think that combined together they paint a bigger and coherent picture, which has more predictive value than each taken separately. We also add some interpretation.
The Eurozone markets will be in doldrums like S&P 500, except for the Germany (DAX) and Great Britain (FTSE). These indexes will outperform the rest of the Europe, but will lag behind the US markets.
The Far East will be down except for the China powerhouse, which will keep growing, but slower. The slowdown in China growth will reflect poorly on their neighbors who supply them with raw materials. Japan (N225) is recovering from the Tsunami effects and will outperform.
Prepare for further food inflation, and industrial raw materials deflation. Base metals and coal will go down. Oil (USO, Brent Crude) and natural gas (XNG) will stay about the same.
Sector wise, hi-tech companies will lead, while the basic industries will lag behind. Banking (BKX) will outshine them all.
US dollar will be up against the EUR and other major currencies except for the Canadian dollar and Japanese Yen, and possibly GBP. Countries with stronger economies will have stronger currencies.
Conclusions and implications:
We see the world coming slowly out of recession in 2013. The crisis in Europe is still not over. The real growth is not on the horizon, but volatility will go down, which points to increased stability.
US markets are largely stabilized, but lack the energy to grow much further, except for the hi-tech, which will lead. We see in today's markets many opportunities in individual stocks and market sectors, both in some beaten up stocks, and in rising hi-tech companies.
This is our forecast for the remainder of 2012 and 2013 based on the data available today. It is a rolling forecast. It is fluid and it's quite possible that in one month the algorithms that monitor the pulse of the market will present an entirely different picture.