This is the title of an article Cascaldi-Garcia et al. (2023) inside Journal of Economic Literature. I summarize some of the key findings below.
definition:
- risk: suitable for situations where the decision maker does not know the outcome, but the probability distribution of the control outcome is known”
- volatility. Often used synonymously with risk, volatility is a statistical measure of changes in observed outcomes
- uncertain. characterized by unknown outcomes and unknown probability distribution
Note the key difference between risk and uncertainty.quote Caballero 2010the article states:
When agents realize that their assumptions about risk are no longer valid and uncertainty conditions apply, their fear of unexpected loss can disrupt financial markets.
Below I summarize four different approaches to measure uncertainty: (i) news-based, (ii) survey-based, (iii) asset-based, and (iv) econometric.
News-based measures.
This approach uses different types of risk as well as mentions in newspapers and other media as quantitative measures of uncertainty.Examples include Baker, Bloom and Davis (2016)The Monetary Policy Uncertainty Index (MPU) is calculated by Hearst, Rodgers and the Sun (2020)and the Trade Policy Uncertainty (TPU) index developed by Caldara et al. (2020). Ahir, Bloom and Furceri (2022) Created the World Uncertainty Index (WUI), which is a GDP-weighted average of country-level uncertainty indices.
Of particular interest is the article by Baker, Bloom, and David that constructs “…the Healthcare EPU Index by searching for articles discussing rising EPUs and terms such as ‘healthcare’, ‘Medicaid’, ‘Medicare’, ‘health insurance’, etc. Articles,” the Affordable Care Act, and Medicare Reform.
The approach is useful because it looks at future risks—particularly geopolitical risks—but may be skewed towards editorial views at major media companies. In addition, causality may be questionable, as suggestions of increased (or decreased) risk for major media entities could influence public perception.
survey-based approach.
This approach requires the individual to consider a variety of different scenarios and assign a probability to each scenario. Uncertainty can then be measured in terms of the standard deviation of the responses in the survey. Ex-ante measurements typically ask respondents for point forecasts (eg, average expectations) of future events (eg, inflation, GDP, sales growth) over a period of time in the future. Aggregating individual responses allowed for an estimate of the dispersion among respondents with respect to point predictions. This accounts for individual uncertainty. Other surveys also ask individuals about the probability of certain events occurring, and so can also be conducted within personal uncertainty. The post-hoc measure of uncertainty compares the deviation of recently released economic data from consensus expectations.
As with all indicators of uncertainty, these indicators have both advantages and disadvantages.
“…survey-based measures can provide an accurate picture of the sector in which uncertainty lies (e.g., firms, households, or traders), economic measures (e.g., employment, spending, policy), and the extent to which uncertainty prevails. However, these Measures tend to be provided less frequently and thus may be outdated relative to news-based or market-based measures.”
Econometric methods.
Econometric methods use data estimation techniques and equate uncertainty with the lack of predictability of aggregation activities. One measure of uncertainty is value at risk (VaR), which is defined as a threshold such that the probability of a particular outcome not exceeding that threshold is equal to the desired level. Probabilities are usually calculated based on quantile regression. More broadly:
“Compared to alternative measures of uncertainty, econometric-based measures have the advantage of being directly based on and guided by statistical inferences, and they reflect the ‘big picture’ in the same sense as news-based measures. However, econometric-based measures of economics are available at lower frequencies and may differ significantly when estimated based on post-revision data versus real-time data,” the article cites a paper Jurado, Ludvigson and Ng (2015)which augments the predictive model with the following factors:
Asset-Based Measures
Historical volatility in return on assets and interest rates is one measure. Asset-based measures tend to reflect the views of market participants who are actively trading in the market for a particular asset and thus may only price risks affecting that particular asset.
“A widely used measure of uncertainty is the VIX, the CBOE Volatility Index, which is calculated using stock index options and measures market participants’ expectations for the volatility of the S&P 500 over the next 30 days.”
The formula for VIX is as follows:
Another measure is realized volatility (RV), which is defined as the proportional sum of the sum of squares of daily returns. RV is considered an improvement over Generalized Autoregressive Conditional Heteroscedasticity (GARCH). Since asset-based metrics typically have large sample sizes and are measured frequently, higher-order moments (e.g., skewness and kurtosis) can also be easily captured.
in conclusion
Unfortunately, this review article doesn’t touch much on health—other than how COVID-19 affects the global economy and Baker et al.’s Healthcare EPU—but it does provide a health overview of risk-related indicators. These empirical approaches to measuring risk using various sources (news, surveys, econometric methods, and asset prices) are readily applicable for healthcare purposes, especially news- and survey-based measures.
