One additional myth suggests a portfolio's exposure to a risk driver can be judged simply by its allocation in terms of weights to a stock or sector. "Looking at your holdings" is not the solution.
To mitigate stock-specific risk, many investors choose to allocate portfolio holdings over as large a selection of stocks as possible and/or simply keep portfolio allocations close to those of the market-capitalization benchmark. However, both practices might lead to overexposure to stock-specific risk factors.
Let us consider a Japanese stock portfolio and try to answer two simple questions about this portfolio.
Question 1: How much is the portfolio exposed to oil price variations?
In order to answer this question, an investor should not run to his desk and count the barrels of oil in the portfolio, the scientific answer to this question consists of computing the portfolio's correlation to the variations of the price of oil.
Question 2: How much is the portfolio exposed to Toyota?
Never answer "the portfolio holds 2.5% of Toyota." If the remaining 97.5% stocks are not correlated to Toyota — your portfolio's exposure to Toyota is actually lower than if you held only 1% in Toyota but the remaining 99% were highly correlated to Toyota. Never forget to wear your correlation glasses!
What matters is not the weight of a stock or a sector in the portfolio, but the portfolio's correlation to a risk driver, whether this driver is the price of oil, Toyota or any other driver.