It may not be enough to understand which ideas and policies could directly improve our society. We also need to grapple with engineering and maintaining them.
In a now-famous article, Patrick Collison and Tyler Cowen proposed the field of Progress Studies. By ‘progress’, they mean ‘the combination of economic, technological, scientific, cultural, and organizational advancement that has transformed our lives and raised standards of living over the past couple of centuries.’ Progress Studies aims ‘to identify effective progress-increasing interventions and the extent to which they are adopted by … institutions … The goal is to treat, not merely to understand.’
I suggest Progress Studies can be divided into two main questions:
What systems would work best to generate progress, i.e. what institutional and other systems should we adopt to create more progress?
Then, even if we manage to answer ‘what’:
How do we actually engineer getting there, including changing laws, institutions and society as necessary?
Most discussion so far has focused on ‘what’. I think that is the easier question. I think we need to focus on the ‘how’ – how to engineer societal change to get more progress in a democracy – and identify potential areas for future research.
I say ‘what’ is easier than ‘how’ because we already have a pretty strong idea  See also Eli Dourado’s post. For each of these points, there is probably at least a minority that would dispute it. That may well be a large part of the problem in fixing them. that:
- Housing is probably some six times more expensive than it needs to be given existing physical technology for building homes.  Hsieh and Moretti (2019) estimated that US growth in productivity per head, controlling for increased education and other variables, would have been one third higher between 1964 and 2009 if San Francisco, New York and San Jose had zoning regulations in line with the median US city, providing improved agglomeration effects.
- Public transit costs four times more to build in the US than in Scandinavia according to Alon Levy, which in turn means much less gets built, increasing congestion. (Transportation and warehousing is 3.1% of US GDP.)
- Healthcare in the United States is less productive  The US spends more per capita on health care than any other country (Helland and Tabarrok (2019), Figure 17) without achieving better outcomes. than in some other countries and less productive than it could be. Medical billing and administration represents a significant, growing, but largely useless percentage of US GDP. For example, the US spends over 6x more than the OECD average on healthcare administration or over $150bn/year.  Cohen and DeLong, Concrete Economics (2016). It is possible that the favorable tax treatment of employer-funded healthcare in the US, coupled with rules increasing free rider problems such as the requirement for hospitals to treat emergency cases without insurance, have made the market less competitive and less efficient. The tort system does not measure consumer preferences for malpractice liability, which may be too large and costly. Healthcare accounts for 16.9% of GDP.
- A significant part of tertiary education is signaling and it therefore could, at least in theory, be made cheaper without losing any of the beneficial effects on society. (9.6% of GDP)
- The banking system and wider financial sector absorb trillions of dollars of implicit and explicit government subsidies every cycle.  For an assessment of the social value of finance to society, see e.g Greenwood and Scharfstein (2013). Subsidies to banks include the discount window; federal deposit guarantees; maturity mismatches allowed in banks that are illegal in most other retail financial services; the too-big-too-fail implicit guarantee; and exclusive access to reserve accounts. Credit provision through the financial sector is probably also inflated by the favorable tax treatment of debt as opposed to equity finance. Having worked in finance for over a decade, I am happy to say that without those subsidies, the sector would probably be smaller and many of the intelligent people employed in negative-sum games in it would probably be better used in other sectors, contributing more on the margin to human welfare and progress. (7.4% of GDP)
- There is no institutional process for making the legal system more efficient and there seems to be much low-hanging fruit. For example, many countries and states have a low-cost and efficient land registration system whereas California has expensive title insurance; similarly other states require expensive lawyers to be involved with home purchases whereas California has a standard form and lawyers are typically not involved.  Moving to an English system requiring the loser to pay a fraction of the winner’s legal costs would also reduce frivolous lawsuits and could be coupled with protection for litigants of limited means. (Legal services were 1.3% of GDP in 2017, and arguably caused far higher friction costs.)
- Even in fields where we still do not clearly know what would be a better system, such as allocation of government funding for scientific research, we do not even have a system for carrying out randomized controlled trials to find out what systems work better, as Azoulay (2012) suggests. If RCTs can be done ethically in medical research when lives are at stake, then they should be possible across many government activities. Technology now permits large firms to carry out millions of A/B tests daily. Government should do the same wherever ethically and politically possible.
Fixing all of those would probably make US standards of living something in the region of 50% higher – more for the UK – and give people far more freedom to be entrepreneurs in innovation clusters to drive faster growth. None of it would require inventing any new physical technology. They are ‘institutional low-hanging fruit.’ Yet they remain unfixed and many or all of them are getting worse. If we knew how in practice to actually get a government to fix these policies and laws, we could apply that to technique to many systems to increase innovation.
Unproductive sectors keep growing
The Baumol effect  See, e.g., Helland and Tabarrok (2019). is the rise of salaries in jobs that have experienced no or low increase of labor productivity, in response to rising salaries in other jobs that have experienced higher labor productivity growth. Over time, sectors with continuing low productivity will take up an ever larger share of the economy. The sectors that it is hard to make more efficient – often because of rent seekers and crony capitalism or because there is a large public element or other public rule that prevents competition and thus constrains innovation – will come to dominate. It becomes increasingly important to fix those problem sectors.
Much historic economic progress has come from producing things with reduced inputs. In the bloated sectors listed above, we urgently need to find ways to do that, by answering both the ‘what’ and ‘how’ questions.
We need to learn how to improve institutions and the law
To change institutions – particularly government institutions in a democracy – is really hard, as Mancur Olson, Walter Scheidel, Daron Acemoglu and others have pointed out. Olson explained how societies will accrete blockages and rent-seeking over time, but we have no institution designed to keep them to a minimum, and collective action problems make removing them difficult.
Most people focused on technology or economics have not done much work on how to achieve political and legal change. With exceptions, effective altruists often assume that systemic change is too hard, but that rests on a flawed zero-sum mental model. Reform is easier than they often think, because the huge inefficiencies of the current situation mean reform can often be designed to be win-win for almost everyone.
Many people from those fields who want to fix society are fascinated by regimes with quasi-dictatorships in small countries like Singapore because they make it easy to neglect the ‘how’ question. It is highly tempting, like the economist looking for the lost banknote under a lamppost, to focus on ‘what’ we want and imagine that we can magically find a dictator to impose the changes we want.
It seems unlikely that the United States or the United Kingdom can move to a Singaporean system of government in the foreseeable future. But there is a range of techniques from different fields to achieve government and institutional change that are not generally well known, and which do not seem to have been much employed to fix the seven problems listed above.
There is some existing political science literature on how to achieve societal change,  See e.g. Saul Alinsky, Rules for Radicals; Bruce Bueno de Mesquita, The Dictator’s Handbook; William Riker‘s ‘heresthetic’; The Letters of Napoleon; the CIA Manual; and biographies of various religious and political leaders. See also Glen Weyl’s Radical Exchange movement. but most political scientists are focused on studying how things are rather than outlining practical methods for how to change them. Institutional economics also partly addresses the question of which institution works best but does not seem focused on the question of how to engineer improvements in the most developed countries. The World Bank and other institutions now do ‘applied political economy’ analysis – finding politically feasible second-best policies that it might be possible to persuade a government and other players in a less developed country to adopt, and working out how to get them to do it. Similarly, the field of policy analysis focuses on designing politically practical reforms. There are other efforts to improve institutional decision making and forecasting.
Despite these efforts, the question remains mostly unanswered: how do we in practice actually engineer that a government fixes these policies and laws? There is astonishingly little focus on that hard question in developed countries, when it may give a higher return on effort than almost any other human endeavor. I believe it should be a core focus of Progress Studies and would like to work with others to summarize the state of the art and identify areas for future research.
Many thanks to Sam Bowman, Matt Clancy, Saloni Dattani, Anton Howes, José Luis Ricón, Shreya Nanda, Ben Southwood, Sam Watling, Stian Westlake, and Nick Whitaker for comments. All errors are mine.