ghariton wrote:When I was young we were taught to look at the probability of a loss and the size of the loss if a loss did happen. Multiply the two together (technically, integrate the various sizes of the loss over the density function of that size of loss) and you get expected value of the loss.
ghariton wrote:But I suspect you (and Bodie) are looking at keeping the two measures separate, not combining them. That raises the question: Is it preferable to incur a small probability of a large loss, or a larger probability of a smaller loss? That will vary with the individual, and so requires thought (which is generally unwelcome).
In the book, Bodie insists on building a plan with no probability of a loss
relative to one's minimal needs (as much as this is humanly feasible). That's why he insists so much on using government-backed inflation-indexed securities to eliminate the risk of not meeting minimal needs.
But we could go beyond the "no-risk" needs
funding and ask your question in the context of funding wants
. Does not meeting one's wants
matter? Yes, it would matter, but not as much as not meeting one's needs
. So, the risk is lower. As for the answer to your question, in that context, it will effectively vary with the investor.
ghariton wrote:There is actually a third dimension, the timing of the loss, if it occurs. For example, losing one's job is a loss, but the impact is greater if it occurs at a time when there is an economic downturn, and jobs are scarce. Similarly, I care less if one of my securities is down when my other securities are up. I care more if there is some negative external event, for example my car just got totalled, and I need money for replacement.
"Uncertainty that matters". Does volatility matter? It might matter a lot if there is a possibility that one needs to make a sizeable withdrawal from a volatile asset during a deep downturn to meet minimal needs, but it might matter less (or not at all) if there's no such possibility.
Your perspective reminds me of William Sharpe's definition of risk: doing badly in bad times
. While I like it a lot, it is not as comprehensive as Bodie's.
ghariton wrote:I come back to my point that there are multiple measures of risk. Different measures are useful in addressing different questions, or for different people addressing the same question. The financial advice industry has much to answer for, in trying to get clients to use a unidimensional scale (volatility, or probability of loss -- how big a loss? -- or other measures).
I agree that there are multiple measures of risk. But, I think that Bodie's "uncertainty that matters" is a definition
, not a measure
of investment risk.