A bottom-up approach doesn’t necessarily translate into higher returns
In 1999-2000, at the height of the internet hoopla, every web business was projected to grow at amazing rates. Totting them all up, it seemed the internet sector would soon grow larger than overall US GDP.
This absurdity reminded one of the 1980s real estate boom in Japan. By late 1989, the aggregation of Japanese real estate projections suggested Tokyo city was more valuable than the state of California (which is around 200 times as large, has 8 times the GDP and contains large metros such as San Francisco, Los Angeles, San Diego as well as Silicon Valley). Similarly, in 2007, US real estate growth projections suggested every individual American would soon become a millionaire home-owner.
In each case, absurd differentials between the top-down and bottom up projections pointed to a bubble. This leads us to an interesting approach. In theory, in well-indexed and highly researched markets like the US and Japan, bottom-up projections exist for most listed companies. And so do macro-economic growth estimates, developed from a top-down perspective. That gives us two pictures of the same market and by extension, the entire economy. If the pictures differ radically, somebody is wrong, and an investor who can identify the “whys” of the discrepancies could make a lot of money.
Top down analysts start with an examination of broad asset classes, national economies and so on. Then it narrows down to attractive sectors and finally, specific stocks. Bottom-up analysts study companies in detail and pay much less attention to broader variables. Even when both approaches lead to common stock picks, the reasons for those picks will be different.
Both methods have pros and cons. The top-down analysts often miss “desert flowers” – businesses that thrive in adverse conditions. Top-down analysts also have a tendency to ignore the finer details that can make all the difference to the profit margins and competitive strengths of specific businesses.
In certain situations that is crucial – when the industry is likely to consolidate, for instance. In industries like PE and VC, or for an entrepreneur, the bottom-up approach is probably far more useful, maybe even essential.
The bottom-up analysts in turn, have a tendency to miss the wood for the trees. Any business operating in a specific geography is exposed to the macro-economic risks of that specific area and a micro focus on the details of the business can badly underestimate those big factors.
Most investors seem prepared to pay more to institutions that displays bottom-up research capacity. However, the extra resources and efforts invested in bottom-up don’t necessarily translate into higher secondary market returns.
Several academic studies suggest that the largest proportion of investment returns depends on picking the right asset class rather than the specific instruments. If the right sectors or even the right national economies can be pinpointed, that’s pretty much good enough. Finding the market leader won’t add a great deal to potential capital gains. Without getting strongly into that debate, one can say that there’s obviously space for both approaches. In bubbly situations, there will be gross differences in projections. But even when the differentials are not so large, an analysis of the differences could be interesting.
For example, let's say that the top-down projections add up to more than the bottom-up estimates. Where is the excess (top-down) growth going to come from? Or conversely, if the bottom-up growth adds up to more than the top-down projections, why is the (bottom-up) growth disappearing?
Exploring such logic, it may be possible to diagnose less obvious shifts in the cycle, even when the differentials are of a reasonable order. We might to able to tell when a stronger than expected recovery is occurring from recession or conversely, the economy is slowing down.
In India, it is tough to test plausible theories and trading models based on this combination-comparison approach. The data is of low-quality, so is the analysis, and there’s an absence of historical data estimates that would help us to back-test.
Nevertheless it is possible to run versions of this comparison. There are discrepancies. Bottom-up estimates suggest the Nifty-Junior group of stocks will grow much quicker than the top-down analysis suggests. This could mean a bubble of sorts.
It could also mean a scenario where top-down analysts are much more pessimistic than warranted. It’s difficult to judge the error margins and the direction of likely error. Regular readers will know that I am pessimistic about the current situation. But there’s certainly room to argue the other side of the case.
