12 The variable FDI Total Share equals the sum of inward and outward FDI stocks divided by U.K.gross value added (again,except for industries that require producers and consumers to co- locate,for which the variable was coded zero).The main concern about this measure is that its denominator covers U.K.activity only,but the numerator covers not just inward FDI into the U.K.but also outward FDI out of the U.K.This mis-match of scope cannot be addressed(in part because the data do not disaggregate host countries for outward FDI),but it likely introduces error in our measurement of the underlying concept of FDI exposure. Our second continuous measure of FDI exposure addresses this concern by including in the numerator only inward FDI.Thus,FDI Inward Share equals inward FDI divided by gross value added (again,except for industries that require producers and consumers to co-locate,for which the variable was coded zero).This measure generates the opposite trade-off:no mis-match of scope,but in theory outward FDI can matter for FDI exposure just as inward FDI does.Inward and outward FDI flows tend to be highly correlated,however,which suggests that on balance FDI Inward Share might be preferred to FDI Total Share. It is important to recognize the level of aggregation for the FDI regressors.Our use of 2-digit industries is dictated by ONS rules on public data dissemination.Theoretically,we could imagine measuring FDI exposure more finely at the level of the respondent's company,rather than at the more-aggregated industry level.?Our specification implicitly assumes that within each industry,all workers perceive FDI threats equally regardless of whether each works for a firm with some FDI.This assumption seems reasonable.We are simply assuming that 7 Of course,this is only a theoretical possibility.Even if we had firm-level FDI data,it would not be usable because the BHPS does not report the respondent's firm.12 The variable FDI Total Share equals the sum of inward and outward FDI stocks divided by U.K. gross value added (again, except for industries that require producers and consumers to co￾locate, for which the variable was coded zero). The main concern about this measure is that its denominator covers U.K. activity only, but the numerator covers not just inward FDI into the U.K. but also outward FDI out of the U.K. This mis-match of scope cannot be addressed (in part because the data do not disaggregate host countries for outward FDI), but it likely introduces error in our measurement of the underlying concept of FDI exposure. Our second continuous measure of FDI exposure addresses this concern by including in the numerator only inward FDI. Thus, FDI Inward Share equals inward FDI divided by gross value added (again, except for industries that require producers and consumers to co-locate, for which the variable was coded zero). This measure generates the opposite trade-off: no mis-match of scope, but in theory outward FDI can matter for FDI exposure just as inward FDI does. Inward and outward FDI flows tend to be highly correlated, however, which suggests that on balance FDI Inward Share might be preferred to FDI Total Share. It is important to recognize the level of aggregation for the FDI regressors. Our use of 2-digit industries is dictated by ONS rules on public data dissemination. Theoretically, we could imagine measuring FDI exposure more finely at the level of the respondent’s company, rather than at the more-aggregated industry level.7 Our specification implicitly assumes that within each industry, all workers perceive FDI threats equally regardless of whether each works for a firm with some FDI. This assumption seems reasonable. We are simply assuming that 7 Of course, this is only a theoretical possibility. Even if we had firm-level FDI data, it would not be usable because the BHPS does not report the respondent’s firm