Goldman Sachs: Should We All Be Macro Stock Screening?

by: Stockopedia

2012 has, of course, yet been another year dominated by Macro, with wild swings in sentiment caused by the European debt crisis and the threat of Middle Eastern conflict. The “risk-on” vs. “risk-off” attitude of the market has been driving most hardened stockpickers nuts. In this context, Goldman Sachs (NYSE:GS) published a very interesting piece back in June this year entitled "Bridging Macro To Micro: GS Top-Down Stock Selection", which looked at the interplay between macro and micro factors in relation to stock-picking. It's focused on Asia, but a lot of the thinking is of general interest.

Goldman's thoughtful approach is essentially systematising the kind of "trouble in Iran -- rising oil prices should drive this stock" anecdotal macro handwaving that you so often see from analysts and market pundits. The aim was to develop a top-down stock picking framework bridging from the macro to the micro by building upon "macro factor mapping, micro-specific comparisons, and business-cycle-based investing".

Bottom-up Screening Isn't Everything...

Unsurprisingly, Goldman Sachs' work found that macro analysis is important to equity returns, even to bottom-up-oriented stock pickers, with factors such as global growth, domestic growth and financial conditions being important return drivers for c. 60% of the market cap of Asian equities. Of course, the problem with macro data is there's so much of it. However, while investors can be overwhelmed by noise, Goldman found certain macro variables to be much more important. Said another way, many macroeconomic variables are highly correlated. They found that stock pickers can simplify macro monitoring by focusing on six factors (discussed below), which explain a significant part of an individual stock’s return variations.

Essentially their top-down stock-picking process involves three steps:

  • What phase of the business cycle are we in and, for that phase, how are the key macro-economic factors positioned?
  • In light of this macro environment, which stocks are expected to do best based on stock specific regressions for each macro factor?
  • Having determined relative macro rankings across the equity market, how are micro factors (i.e. valuation, fundamentals and technical indicators) likely to impact expected returns?

The Business Cycle And Macro Factors

As explained in more detail in their paper "Global Economics Paper No: 214, Acceleration Matters: Asset Returns and the Business Cycle", using their proprietary "Global Leading Indicator" (GLI), the GS team define four distinct phases of the business cycle by looking at the interaction of GLI growth/momentum with GLI acceleration (changes in momentum growth). The GLI looks to be essentially Goldman Sachs' souped-up version of the OECD Leading Indicator. The four phases are: 1) expansion (i.e., positive growth and positive acceleration), 2) slowdown (positive growth and negative acceleration), 3) contraction (negative growth and negative acceleration) and 4) recovery (negative growth and positive acceleration).

Depending on which phase we are in, this will then correspond to a certain range of values for the following six key macro-factors:

  1. Market risk (i.e., the "beta" versus the local market index)
  2. Domestic growth - Goldman use something called the Current Activity Index (NYSE:CAI) to estimate the trend rate of growth in GDP.
  3. Domestic policy/liquidity - This assesses how accomodative monetary policy is measuring using their Financial Conditions Index (FCI) - discussed in depth here.
  4. Domestic inflation (as measured by the CPI)
  5. Oil prices (as measured by the WTI)
  6. Global growth (as measured by the GLI)

Using regression techniques, Goldman was able to establish linkages between these macro factors and each stock's expected returns. As you'd expected, there were varying degrees of macro explanatory power for each stock (as represented by the R-squared of the regression). For example, Goldman found TATA Consultancy Services has tended to move very closely with GLI momentum (a 1% GLI increase apparently corresponded to a 15% share price increase), which they attributed to TCS’s business concentration in global developed markets. In the case of Industrial and Commercial Bank of China, they found that 74% of share price variations can be explained by four factors: its high market beta (presumably reflected given the centrality of China concerns to the market return), inflation changes (because of the knock-on effects for loan pricing), financial conditions (because this impacts loan growth and NPL risks) and, finally, global growth (although, interestingly, this was not as sensitive as one might expect for a Chinese bank).

Not Forgetting The Micro — Valuation, Fundamentals And TA

While this analysis partly explain returns variations, they found that no stock had close to a 100% fit in such a macro-factor regression model. Unsurprisingly, there are other factors at play that drive stock returns. Some stocks are driven more by the macro, others more by the micro (i.e., EPS growth and valuation levels). In the case of a company like Samsung (OTC:SSNLF), they found very low correlation with any major macro-factors.

To address this, Goldmans then overlaid micro considerations such as a) valuation, b) micro-fundamentals, and c) technical entry levels to derive a micro Z-score:

A. Valuation Parameters: Interestingly, they chose three main valuation factors: 1) Forward Price Earnings, 2) Trailing Dividend Yield, and 3) Trailing Price to Book. While they found that Price to Cashflow has strong predictive power on forward returns at a market level, they excluded it for stock picking given its short and unstable time series at an individual stock level.

B. Micro-Fundamentals: They then look at the changes in consensus earnings or earnings expectations to assess a stock’s micro dynamics, specifically: 1. Earnings revision (magnitude), 2. Earnings sentiment (breadth) and 3. Consensus target price change. Although there is some signal in recommendation change, one wonders why they didn't opt for the Piotroski F-Score as a better signal for the trend in fundamental health.

C. Technical Indicators: Finally, their approach looks at technical indicators to form a view on short-term entry levels. They group TA indicators into four categories -- momentum, volatility, trend and volume -- and choose to focus on the following three indicators as being most commonly used: 1. RSI (for momentum) which we discuss here, 2. Bollinger bands (for volatility), 3. Moving Average Convergence/Divergence (discussed here). Interestingly, they disregarded volume-based indicators on the basis that they require a time-series perspective to form trading signals and an absolute number does not tell much. You can market screen based on RSI and MACD using our new Chart Signals package -- we'll be adding top down Bollinger Band screening to this shortly.

Does It Work?

Over 10 years of backtesting, Goldman's “macro + micro” portfolio showed an impressive 11.4% average annualized outperformance (326% versus 165% for the benchmark on a market-cap-weighted basis). One clear concern -- acknowledged by the authors -- is the extent to which the relationships found were just data-mining (e.g., were they identifying historical patterns as factors, when in fact they may just be coincidence?). Due to limited data availablity for Asian stocks, the backtest was conducted on an in-sample basis, which is less than ideal from a statistical standpoint.

A final issue is that this kind of macro screening relies on having a decent algorithm and a set of macro regression factors for every stock in the market, which is not something you tend to just find lying around, unlike fundamental factors, which are conveniently listed in the annual report (more or less). Still, it's interesting to see it implemented, and it wouldn't be beyond the wit of a sufficiently motivated investor to generate something similar given, say, 10 years of stock prices and the corresponding macro-economic data points.

Further Reading

  • GS Global Strategy Paper: No. 3: AsiaPac Valuation: What works, and when
  • GS Portfolio Strategy: “Corporate Asia” Earning: from Macro to Micro
  • GS Portfolio Strategy: What macro indicators matter for Asian markets?
  • Global Economics Paper No: 214: Acceleration Matters: Asset Returns and the Business Cycle