As part of our own globalization effort, we analyzed McKinsey’s geographic footprint. We found the analyses interesting so we share them broadly. Technically, they also show the usefulness of probit regression and the power of non-obvious data sources.

McKinsey is a rational firm. Thus, its office locations should indicate where business opportunities lie. What interests us is to create a measure of how strong the various McKinsey locations are. Strong in the sense of whether the underlying market is solid.

To quantify this, and only having public data at hand, we built a probit model which estimates the probability of having an office. A probit model has a binary outcome as dependent variable. In our case, “there is an office in the city” = 1, or “there is no office in the city” = 0. We have 997 cities in the dataset, of which 107 have McKinsey offices.

We found four independent variables that explain office probability: a) city population, b) city GDP/capita, c) if the city is a capital, and d) if the city is an international finance or multinational headquarters center.*

The graph above shows the probability of a city having a McKinsey office as a function of city GDP (keeping other explanatory variables constant). The red ovals show interesting outliers.

- Two cities have over 80% probability of having an office, yet do not have one: Kuwait City and Tehran. They are thus likely new offices (dependent on political considerations). Their probabilities are higher than Osaka’s (77%), which used to have a McKinsey office

- Fourteen cities have offices with less than a 20% probability**. These look like marginal offices (but in the case of Africa, they may reflect a future growth potential)

We now turn to intellectual contributions by office. Which McKinsey offices contribute the most to the firm’s intellectual capital?

We reviewed five years of McKinsey Quarterly articles to see where the authors came from. The sample has 131 articles with 894 pages. The leading contributors are London (95 pages), Shanghai (66) and Chicago (64). Around half the offices made contributions.

##### CLICK ON THE MAP TO ENLARGE IT

**The red circles’ areas are proportional to pages published. The blue triangles are offices with no contributions**

Contributions are led by the US, northwestern Europe and China. That China has such a strong position is remarkable.

The dearth of intellectual contributions from Japan, Latin America, southern Europe and Scandinavia is interesting. One may wonder if clients are well served by those offices.

Finally, we looked at cross-regional collaboration. Slightly more than 50% of articles with more than one author had such collaboration. This is much higher than for academic papers, and probably takes a conscious effort on the part of McKinsey.

In conclusion, the combination of high probability of an office existing with strong intellectual contributions to us signify robust offices. The analyses also show the power of using probit on limited datasets to extract strategic insights.

— — — — —

* The equation in Stata is: probit office lnpop lngdp_cap capital center

** Staffan Canback, our managing director, founded the Gothenburg office–a seemingly irrational initiative