Generating Alpha at OPM

"It ain't what you don't know that gets you into trouble. It's what you know for sure that just ain't so." – Mark Twain[1]

For generations, policymaking has been dominated by lawyers. In an ecosystem naturally selecting for nuanced legal interpretations and the ability to navigate byzantine processes, lawyers were the fittest. And, despite that, some of my best friends are in fact lawyers!

Today, AI is leveling the playing field for those with expertise in other disciplines. For policymakers, legal issues that would have taken months for the uninitiated research can now be understood with a simple prompt. This new superpower will enable those in broad-ranging disciplines to develop policy, and biased as I may be, I would argue investing could be the new generalist skillset best suited.  

Alpha

The modus operandi of any professional investor is to generate what is known in the investment industry as "alpha," which is the excess return an investment generates relative to an appropriate benchmark. For example, many hedge funds try (and fail) to outperform the S&P 500 index, which can be earned with zero investment skill. Therefore, the value that an investor adds is measured by the return in excess to this benchmark (on a risk-adjusted basis).

So how does generating alpha relate to policymaking? An investor must first understand what the consensus view is and then determine whether the consensus is right or wrong. The greatest opportunities are found when there is a strong consensus that is also wrong. Having a both contrarian and correct view is the key to generating alpha.  OPM Director Scott Kupor’s former a16z partner Chris Dixon likes to say that great investors invest in good ideas that look like bad ideas. The trick of course is not being confused by how things may at first appear, but rather homing in on their differentiation, which may at first make them look like silly or bad ideas.

The same is true with effective policymaking. Many of the most intractable policy problems are difficult for good reasons, and the consensus correctly understands what that reason is. However, the best opportunities to fix issues arise if a policymaker understands what the consensus is but also believes the issues can be addressed with some contrarian thinking.

Alpha Policymaking Applied: Tech Force

The federal government has previously tried, with varying levels of success, to recruit exceptional technologists into government. When designing a new program, with hopes that it would be the most successful, we had to first figure out the consensus view. The consensus view was generally pessimistic and contained a few key convictions: (1) the government cannot compensate well enough to attract top-tier engineers, (2) technologists are not interested in government work even at higher compensation levels, (3) a centralized hiring model would be needed, (4) the private sector won't want to engage in the program, and (5) even if successful, in the age of AI, engineers won't be necessary anymore. We thought all these conclusions were likely wrong, and therefore, if we were right, we could generate alpha. That is, we could design a program that beat expectations and overachieved.

Compensation is Not a Panacea

Compensation is typically the first challenge cited in government recruiting. But too often compensation is used as an excuse to cover up other failings in the recruitment process. For example, even if the government could offer higher salaries, it would need to ensure it targeted the right populations, accelerated the time to hire, and offered compelling career opportunities to get the best people. Compensation is not a panacea.

So for Tech Force, we first targeted our recruiting efforts on early-career engineers - this is where the pay gap with the private sector is the smallest and where each dollar spent on salary could offer the highest return. Each year, millions of people early in their careers take out loans and pay to gain educational experiences and build their career, so there should be no doubt that if the government job is compelling enough, young technologists are willing to get paid to do it.

We also recognized that compensation is the first derivative of appropriate leveling of skillsets. A huge barrier to hiring highly qualified, yet early career talent has been arcane leveling rules that prohibit the government from assessing people based on merit, but rather rely on proxies for merit – most fundamentally, years of experience. We recognized that we needed to eliminate these proxies to properly level an early-career engineer: If you can perform at the level of a GS-14, we should pay you at that level, independent of whether you have 10+ years of work experience. Proper leveling really compresses the perceived pay gap in a substantial way.

Second, we attempted to run a personal recruiting process that moved a person from job application to offer as fast as possible. While there is room for improvement, it is improving!

Third, we made a pitch directly to engineers that the work in government is unique and meaningful. The scale and complexity can only be found in government, and you get to solve problems on behalf of every single American. No start-up can make a similar pitch.

Technologists are Interested in Government

Tech Force has received massive interest from candidates who want to put their skills to work in government. Instead of building another consumer application, they yearn to build something that tangibly improves the country. Thus far, Tech Force's talent network has grown to over 50,000 members and we've received over 10,000 job applications. There is a groundswell of interest from patriotic and talented people who have been looking for such an opportunity.

Centralized Recruiting. De-centralized Hiring

Several previous hiring initiatives hired engineers centrally and then would deploy them out to agencies as needed. For Tech Force, OPM has centrally handled recruiting and assessments to certify eligible candidates to be hired by agencies but has left the actual hiring to be completed at the agency level.

This has two key virtues: one, the program can scale much better because it can immediately access the budget of every single agency and two, it creates accountability at the agency level – the agency is responsible for making the hire, ensuring the individual is deployed in the best way, and is responsible for managing the individual's performance. This discipline mechanism will result in improved performance and experiences for the engineers. OPM will leverage this model in many other fields as we leverage cross-government job announcements to centralize recruiting and de-centralize hiring.

Industry Partnership

No previous initiatives attempted to partner with the private sector. Tech Force recognized that early career engineers want to create career optionality and receive validation that the private sector values the experiences they will gain in government. Tech Force now has over 40 industry partners who recognize the value of civic tech experience. These partners see a clear value proposition in supporting this ecosystem of talent with programming and training and then being able to recruit these individuals after their terms in government.

In addition, our partnership with private industry changes the career trajectory for our participants. We are no longer trying to market a 40-year federal career to early career candidates who don’t think in those terms. Rather, our value prop is simple – spend a few years in the government, work on the coolest and most complex problems in the world, meet a bunch of like-minded individuals, and then if you decide to go to the private sector, we’ll help you make that transition. This is how modern careers will work; the government needs to reform its message to align with that.

Demand for Engineers

Over the last few years, as AI has rapidly improved, many prognosticators have predicted the extinction of software engineers. While no one can predict the future, it appears at least for now, that the death of software engineers has been greatly exaggerated. For many in the private sector and certainly in government, these skills are actually even more valuable. While AI has changed workflows, we've made a bet that software engineers will be needed in government for many years to come. To the extent other employers have disagreed, we're happy to keep benefiting from our initial contrarian bet – send them our way!

Compounding Alpha

Our approach to Tech Force shows the value of contrarian thinking to solve problems and generate alpha. In the age of AI, using this lens will enable people from diverse disciplines to solve policy problems with new ideas. Under President Trump's leadership, agencies across government have the mandate to build and implement creative policy solutions. At OPM we will continue to "ask why," use a contrarian lens to generate alpha, and then consolidate and compound these wins to do just that.  

[1] Widely attributed to Mark Twain but the underlying aphorism has been best traced to the 19^th century American humorist Josh Billings.