Kludge and how think tanks and policy wonks make it worse

Think tank scholars and policy wonks strive to be both practical and clever. Being practical means proposing policies that have a good chance of getting taken up by government in the short term. And being clever means policies that generate big benefits at little or no cost. But according to American political scientist Steven Teles, the short term benefits of practicality and cleverness have long term costs.

Americans argue endlessly about the size of government. But a more pressing problem is the complexity and incoherence of government programs. According to Teles, American government has succumbed to kludgeocracy:

A "kludge" is defined by the Oxford English Dictionary as "an ill-assorted collection of parts assembled to fulfill a particular purpose…a clumsy but temporarily effective solution to a particular fault or problem." The term comes out of the world of computer programming, where a kludge is an inelegant patch put in place to solve an unexpected problem and designed to be backward-compatible with the rest of an existing system. When you add up enough kludges, you get a very complicated program that has no clear organizing principle, is exceedingly difficult to understand, and is subject to crashes. Any user of Microsoft Windows will immediately grasp the concept.

And so too will any close observer of US public administration. Speaking to the Washington Post’s Ezra Klein, Teles said: "A program or policy qualifies as a kludge if the fundamental policy mechanism is substantially more complicated than the problem it is trying to solve dictates. In general, it is a "kludge" because it builds upon, rather than supersedes, the policies that came before it.”

For think tank scholars and policy wonks, ‘practicality’ often means proposing measures that sit on top of existing programs rather than arguing for systemic reform. While this gets results in the short term, in the long term it leads to a build up of kludge.

According to Teles, policy wonks and think tank scholars have "a preference for clever or innovative policy mechanisms; relatively simple, direct uses of governmental brute force are just not as interesting." One result is privatised programs that rely on a complicated web of tax breaks, subsidies and regulation.

In political debates conservatives and liberals argue about whether services should be performed through the market or through the government. But too often, governments don’t choose one or the other. "We’ve tried to have both," says Teles," and in having both, we lose a number of the advantages of either. In most cases, ‘privatization’ does not really mean that a function has been given back to the market. It means that we have a highly subsidized, regulated, sometimes monopolized activity in which there is private ownership but a high degree of public control."

A first step towards clearing away kludge is to help ordinary citizens see how the problem affects them. Once people understand the problem, they’ll see how seemingly unconnected complaints have a common cause. According to Teles, raising awareness is:

… the work of public intellectuals, bloggers, researchers, and entrepreneurial politicians. Only the shapers of public debate can help the public recognize that the source of the insider dealing and special interest politics they detest is the policy complexity that their own ideological incoherence helps to create.

Find out more:

Steven M. Teles. ‘Kludgeocracy: The American Way of Policy‘, December 10, 2012, New America Foundation

Matt Steinglass. ‘Health-care reform: America’s kludgiest kludgeocracy‘, September 27 2013, The Economist

Reihan Salam. ‘Steven Teles’s Really Important Essay on Kludgeocracy‘, December 10 2012, National Review

Kevin Drum. ‘The High Cost of Rube Goldberg Policymaking‘, December 11 2012, Mother Jones

This entry was posted in Political theory, Politics - international. Bookmark the permalink.

24 Responses to Kludge and how think tanks and policy wonks make it worse

  1. David Walker says:

    “Americans argue endlessly about the size of government. But a more pressing problem is the complexity and incoherence of government programs.”

    Great quote, great thought. A few quick responses:

    * One of the worst things about complexity and coherence is that they are so hard to measure. And without measurement, it’s harder for people to see the problem. I don’t know how to measure complexity and coherence, but it would be worth trying a system to have the public service at least report on it.

    * Regulatory impact statements – which were partly designed as an anti-kludge measure, and which Teles essentially suggests for the US system – have been generally ineffectual, and in my very limited experience have not made much effort to point out rising kludge.

    * Big-picture investigations and independent evaluations have been more successful. The Productivity Commission has done some good work pointing out kludges in various policy areas. The Henry Review of tax was an attempt at a path to de-kludging the tax system. Performance audits by auditors-general can also play a role here, though they currently receive less attention than they merit.

    * Teles would like to see “fewer, but larger and more effective pieces of major legislation”, and suggests super-majorities might help get it. Maybe, maybe not. They might also have the opposite effect, by encouraging parliamentary wheeling and dealing.

    * Teles makes an interesting argument for pushing legislative micro-design down to the agencies that will run the processes to make the legislation work. This has actually been a very lively dialogue in certain parts of Australian government – notably the relationship between the Treasury and the ATO on design of tax legislation.

    * Teles makes a useful analogy between the government and Microsoft Windows. Windows, like governments, is complex in part because of the continuous need for backward compatibility. You can still run DOS programs on some versions of Windows 8, and today’s superannuation rules still account for 1970 legal arrangements. “Backward compatibility” is an important asset of many government systems.

    * Finally, one more software analogy. Linux and iOS work well in part because of good underlying architecture. (The US system may have a real problem here.) But Linux and iOS also work well because they are to some extent relatively recent clean-sheet exercises. Programmers love such exercises; they’re a lot more fun than cleaning up (refactoring) existing code. When you start to see the same problem in government, there’s a similar temptation to opt for clean-sheet thinking. And occasionally they are necessary. But in most circumstances there’s a lot to be said for refactoring in programming, and there’s a lot to be said for smart incremental clean-ups to our current system.

    • Nicholas Gruen says:

      I might try to post on this myself, but for now count me in with David “in most circumstances there’s . . . a lot to be said for smart incremental clean-ups to our current system.”

      And the Henry Review was a train wreck from the start. I mean really? An eighteen month ‘clean sheet of paper’ review to report six months before an election. Oh – and that clean sheet of paper? Well it was clean ish. Couldn’t involve the GST. And in the midst of a tiny scare campaign the Treasurer told us that he liked dividend imputation. And so it went. And when it landed surprise, surprise the eyes got picked out of it.

      There’s nothing so naïve as a group of powerful people all coming up with a Grand Plan.

  2. Hi David,

    Nice commentary. Just to pick up on one point:

    Teles makes an interesting argument for pushing legislative micro-design down to the agencies that will run the processes to make the legislation work.

    There certainly have been many occasions where legislation is, in practice, unworkable to implement actual processes for. But I’m reluctant to hand over control of the legislative detail to the implementing agency, since they will inevitably move the responsibility for handling complexity onto citizens.

    Rather, I would prefer to see legislation that clearly outlines objectives, circumscribed behavior, and public-facing metrics to allow people to assess whether the desired objectives are being achieved.

    Notwithstanding the tendency to game metrics, as long as no explicit targets are ascribed they should still usefully gauge how well processes have been implemented.

  3. Patrick says:

    Teles is doubtlessly thinking of the US model, in which agencies already do a significant amount of the relevant micro-legislation: The IRS and the SEC write their own laws to thousands of pages worth, as does the EPA, the FTC, the FDA etc.

    I’m not sure how much more he wants to push down to these guys!

    Personally, I prefer our system which has a greater degree of ‘micro’ –> APRA writes prudential standards, the ATO provides rulings but these are only quasi-law (they aren’t even regulatory instruments), the AASB provides accounting standards, etc.

    This is because one product of the US system (in my limited to almost zero experience) is that Congress effectively delegates its power to so broadly (the SEC shall make provision for the regulation of banks, basically!) that ultimately it is an either substantially captured or under-resourced or inappropriately adversarial agency which makes the critical allocation of resources, consumer/citizen rights and due process questions which should in fact be made by elected representatives.

  4. Alan says:

    The US has a huge problem with secondary legislation at the federal level because the supreme court ruled, in Immigration and Naturalization Service v. Chadha, that congress cannot provide for secondary laws to be disallowed, or even subjected to a congressional veto. Secondary legislation exceeds primary legislation (acts of parliament) in volume by many times in all modern governments. It would be hard to find an Australian act that does not authorise the governor-general,the minister or the agency head to make secondary legislation.

    What happens instead is a semi-judicial process that is both kludge-friendly and a rentier’s picnic.

  5. Patrick says:

    I’m not sure that any of that is really right.

    I’m not sure that case means what you take it to. It simply provides, as I read it, that as legislation must be passed by both houses, one house acting alone can’t veto it. Which is also the position in Australia, as it happens.

    Also:

    Secondary legislation exceeds primary legislation (acts of parliament) in volume by many times in all modern governments.

    This would be the case in the US, but I’m pretty sure it wouldn’t in Australia. And in Australia delegated legislation usually has to be presented to Parliament for approval anyway, as far as I can recall.

    • derrida derider says:

      Yes, that’s all correct. Mind you, the Australian ratio between primary and secondary legislation (“delegated legislation” in our jargon) is in part driven by the great volume of primary legislation.

      Federal delegated legislation is “disallowable” – the parliament does not have to approve them, but can veto them. Interestingly, either House can do the veto – which is one reason control of the Senate really matters.

  6. Alan says:

    An example from NSW where 409 subordinate laws were passed while 99 primary laws were considered by the parliament in 2007.

    Chadha has been consistently read as preventing disallowance by one or both houses of congress (negative orders) or even a requirement that a law be presented to congress and not enter into force until passed by both houses (positive orders). Being sure of things is fine. Being accurate about them is better.

    • Patrick says:

      Ok, NSW passed a lot of subordinate legislation. I admit that I don’t know much about NSW, and planning laws in particular might distort those numbers.

      But if the Cth is anything to go by, the vast majority of those would have been very short, whilst a great number of the actual Acts would have been very long.

      As for Chaudra I still don’t get what you think is the big deal. According to you (and I’m willing to believe you based on my very quick reading of it), it basically says, as far as I can tell, that neither or both houses of Congress can circumvent the Constitutional requirements for passing laws under guise of ‘vetoing’ existing laws.

      That isn’t the case in Australia for obvious reasons, but I don’t see it as any real handicap for America. All law is actually originated in one of, and passed by both, houses of Congress. If they were concerned that the legislature would run around making delegated legislation that they didn’t like they could simply legislate themselves, or provide a sunset/approval clause, etc. It isn’t in any meaningful sense a restriction on the legislative supervision of delegated legislation – the meaningful restrictions are capability and the sheer volume of delegated legislation that the US system does, under Congress’ express authorisation, create.

      • Alan says:

        The stats are against you for the Commonwealth as well. Australia, incidentally, has a reputation for using primary legislation much more than comparable countries. Britain is famous for leaving everyone to secondary legislation.

        As fro Chadha, I am unclear that you understand the distinction between primary and secondary legislation. The legislature does not, by its nature, make delegated legislation as your post implies. The reason Chadha has devastated democratic control of secondary legislation in the US is that any congressional action to disallow a secondary law must pass the president’s veto, like all primary laws in their system.

        Allowing Congress to supervise delegated legislation only with the executive’s consent does not strike me as a miracle of institutional design. It’s hard to imagine a more kludge-freidnly system, than making secondary legislation a double process of semi-judicial hearings conducted exclusively by and with lawyers, and then judicial review in the courts.

        • Patrick says:

          Are you just defending your position or do you have particularly strong ‘priors’ that make it hard for you to be open to a different position??

          Your second sentence sounds accurate to me, but it contradicts your first sentence. As does the ‘data’ you helpfully linked to!!

          First, the Cth has more pages of primary Acts than of regulations, and that doesn’t take into account the number of pages of amendments passed to primary Acts each year which I assure you vastly outweigh the amendments to regulatory instruments.

          Second, your prime example was NSW, but they appear to have passed approximately 4.5 times as much primary legislation as delegated legislation. No State comes even close to “more” delegated legislation on a page count (which is surely more relevant than number of instruments since a huge number of those instruments simply name a specific entity or group of entities or designate planning zones or define a particular unit or number (eg penalty units).

          Finally, back to the US where I completely expect that they do pass far more delegated legislation. To be perfectly clear: the Legislature controls delegated legislation to almost the exact degree it chooses to. It is constrained only by three factors in the US:
          – it’s time;
          – that it doesn’t appoint the regulatory agency staff; and
          – that it has to pass a new law (in crude effect, please just live with that little inaccuracy) to overturn instruments once made.

          They can however define
          – the scope of the delegated legislation;
          – which agency is to make it;
          – what factors the courts must take into account in considering the validity of the regulations;
          – what factors the regulators must take into account;
          – whether there is to be any delegated legislation at all;
          – the legal effect of the instruments to be made;
          – the period of validity of the instruments to be made;
          – etc ad nauseam.

          So what is your “democratic deficit” problem?

        • Patrick says:

          Do you simply mean to assert that modern government is inherently anti-democratic in its scope and size?

        • Alan says:

          The democratic deficit is that any primary law in the US is subject to presidential veto. A veto requires a 2/3 vote in both houses.

          Since secondary laws are made by the president or officials appointed by the president (and in most cases answerable to the president) that strikes me as a major problem.

        • Patrick says:

          Fair enough, but I don’t think there’s an intrinsic democratic deficit in what you describe. There might be a de facto one arising from lack of Congressional ability or willingness to draft legislation properly, but that’s closer to my last suggestion than to anything you’ve said.

        • Alan says:

          Patrick

          The US has two problems, the breadth of secondary legislation passed by the congress and the impossibility of congressional review. The second is the large volume of national security legislation made by the president as commander-in-chief. Examples include the whole of the occupation code which applied in Iraq, the regime of imprisonment and detention in the War on Terror and a number of other things that really belong in acts of congress.

    • Ken Parish says:

      Dear Alan

      Patrick is right about this. Chadha merely held that a single house veto of delegated legislation to be unconstitutional. The US now has a bicameral system for review of delegated/subordinate legislation. See the Congressional Review Act 1996 and this article starting at page 15.

      • Alan says:

        Deasr ken

        The US supreme court itself said in Chadha that it applied to a bicameral legislative veto as well as a unicameral legislative veto. I can cite the language if you will, but I am aware this is getting to be a fairly technical discussion and one that that doesn’t much bear on the original issue.

        The Congressional Review Act does not, because it cannot in terms of Chadha, set up a legislative veto. It sets up an expedited process in the senate (but not the house) to pass a primary law that still requires the president’s signature. The Act supports my argument and is merely another case of the tyranny of the Presentment Clause.

        • Ken Parish says:

          SCOTUS said no such thing. I presume you are referring to this passage from Burger J’s judgment:

          “The legislative character of the one-House veto in these cases is confirmed by the character of the congressional action it supplants. Neither the House of Representatives nor the Senate contends that, absent the veto provision in 244(c)(2), either of them, or both of them acting together, could effectively require the Attorney General to deport an alien …”

          It is simply saying that even both houses acting together could not validly disallow regulations absent enabling legislation setting up a framework for them to do so. The Congressional Review Act 1996 established that framework. It IS a mechanism for review and disallowance of subordinate legislation by both Houses of Congress.

          Of course Congressional disallowance motions are subject to presidential veto, but so too is primary legislation. That is the American system. It is no more fundamentally anti-democratic than having a Westminster system where the executive government dominates the parliament because it effectively has the numbers at least in the Lower House. Westminster arrangements effectively ensure that an executive government usually won’t be foisted with administering legislation or regulations with which it radically disagrees. The US system involves more complete separation of powers but achieves a similar outcome by means of presentment/presidential veto. Recent history suggests our arrangements are currently working more smoothly than the US, but that might be more due to immediate political cultural issues (Tea Party loonies) that defective governance design.

          Incidentally, Obama appears only to have exercised the Presidential veto on two occasions and GW Bush on only 12, and none appears to have been a veto of Congressional disallowance of delegated legislation. Thus the danger to which you refer seems to be more apparent than real. OTOH Congressional disallowance/disapproval is also uncommon, no doubt due to the requirement for both Houses to approve the motion.

          This Harvard Law Review article provides a rather more sophisticated analysis of the CRA than is possible in a blog comment box.

          Incidentally, my experience as a member of the NT Legislative Assembly Subordinate Legislation Committee over some years taught me that the Westminster system does not provide parliament with any more effective oversight of regulations etc than the US system, at least in states/territories with a unicameral legislature.. The government of the day always has the numbers, and also ensures that the Committee does not have the resources to effectively exercise review oversight of delegated legislation.

          The equivalent federal parliament committee is the Senate Standing Committee on Regulations and Ordinances. It is somewhat better resourced than its NT equivalent and currently has 3 Liberal and 3 Labor members, and thus is much less a helpless captive of the executive government than its NT equivalent.

          However its remit does not focus on economic efficiency/de-“kludging” but rather on human rights and similar issues. I agree with a previous commenter that giving the Productivity Commission a much more prominent role in this area would be a good idea.

        • Alan says:

          We have read the same article. For me the most salient part of the Harvard Law Review article appears in the concluding paragraph:

          But this Note has shown that there is a bit more to the CRA. Two facts are most salient. First, Congress chose to create the CRA. Second, Congress has never used it, except in the case of the ergonomics rule, which may be sui generis.

          A single use in 17 years cannot seriously be claimed as a generalised system of legislative veto. An equally persuasive analysis of Chadha’s impact concludes:

          The predictable and inevitable result of Chadhais a system of lawmaking that is now more convoluted, cumbersome, and covert than before. Finding the Court’s doctrine incompatible with effective government, the elected branches have searched for techniques that revive the understandings in place before 1983. In many cases, the Court’s decision simply drives underground a set of legislative and committee vetoes that formerly operated in plain sight. In one form or another, legislative vetoes will remain an important mechanism for reconciling legislative and executive interests. The executive branch wants to retain access to discretionary authority; Congress wants to control some of those discretionary decisions without having to pass another public law. Such accommodations are better fashioned by committees and agency officials than by judicial decisions, especially the broadly crafted rules announced in Chadha.

          Patrick initially argued that Chahda only addressed the unicameral legislative veto. I argued other wise. The court says:

          8. The congressional veto provision in § 244(c)(2) is unconstitutional. Pp. 944-959.

          (a) The prescription for legislative action in Art. I, § 1 — requiring all legislative powers to be vested in a Congress consisting of a Senate and a House of Representatives — and § 7 — requiring every bill passed by the House and Senate, before becoming law, to be presented to the President, and, if he disapproves, to be repassed by two-thirds of the Senate and House — represents the Framers’ decision that the legislative power of the Federal Government be exercised in accord with a single, finely wrought and exhaustively considered procedure. This procedure is an integral part of the constitutional design for the separation of powers. Pp. 944-951.

          (b) Here, the action taken by the House pursuant to § 244(c)(2) was essentially legislative in purpose and effect, and thus was subject to the procedural requirements of Art. I, § 7, for legislative action: passage by a majority of both Houses and presentation to the President.

          The Presentment Clauses certainly mandate bicameralism. They also mandate the president;s veto. If the decision turns on the unicameral legislative veto, why mention the veto requirement? And what source of authority in the text of the constitution itself, by your argument, would justify a system of bicameral disallowance that the Congressional Review Act itself makes no attempt to establish.

          Disallowance subject to veto by the executive is not disallowance. It is subject to the exact same weakness identified in our house of representatives in the 1930s and indeed by you in the NT legislative assembly right now:

          Maurice Blackburn, later a Labor member of the House of Representatives, had explicitly contended in 1930 that:

          the House of Representatives is not likely to do that work well, or, in fact, to do it at all. Upon its vote turns the fate of the ministry. The regulation is made by the ministry, and a proposal for its disallowance would certainly be treated as a vote of want of confidence, and would be tested on party lines. No ministry depends on the vote of the Senate and it is quite likely that in that chamber a regulation would be considered on its merits…. (Evidence to the 1929 Select Committee, PP S1/1929-30, p. 23.)

        • Ken Parish says:

          Alan

          If you judge the efficacy of a regulatory review system by the frequency of disallowance in the Australian system isn’t very effective either. I have only checked back over this year and last, but it would appear that no delegated legislation has been disallowed at Commonwealth level in either 2012 or 2013.

        • Alan says:

          Ken

          The most alarming statistic of the congressionalr eview story is not the single disallowance by reapeal in the lifefo the CRA.it is that the regualtion dislalowed, the OSHA ergonomics regualtion, took 10 years to enact. Moreover much of the reason for that extraordinarily long enactment time was various measures which in theory militate against kludge, such as regulatory impacts.

          The US has added so many anti-kludge measures to tis system of rule-making that the act of rule-making has become impossible.

          In light of that I would prefer the review of secondary legislation to stay with the Senate committee and for that committee to have its remit and resources expanded. Generally I favour democratic parliamentary committees over econocrats gazing at their navels.

        • Patrick says:

          I think we’ve moved a long way from the starting point. It appears that we now all agree that there isn’t an institutional or legal democratic deficit but possibly a practical or cultural issue with review of delegated legislation.

          I’m not sure there is a real issue there either. In my experience, which admittedly is quite limited, delegated legislation in Australia is not such hefty stuff – as I noted above, there are reams of designations of particular bodies as a charity, particular organisations as terrorists or foreign aid recipients or disaster-stricken, or designation of amounts and dates. One wouldn’t expect this stuff to be overturned regularly!

          It remains the case that in Australia the vast majority of legislating is done by Parliament.

          In the US this doesn’t seem to hold. Again as I observed above, for reasons of its own the US Congress delegates incredibly broad powers to the SEC, Federal Reserve, IRS, EPA, FDA, FTC and a million other acronyms. The delegated legislation passed by these bodies is a real issue because it determines very substantive legal questions of rights and duties, and its complexity and length makes it utterly beyond effective Congressional review.

  7. Tel says:

    The theory of tied legislation is a Pareto optimisation step.

    If enough baggage is tied to a change, then everyone feels they got something out of it, so therefore we can say that all parties improved their position. Essentially, people who were disadvantaged by the primary legislation got something, by being paid off to accept their disadvantage in order for everyone else to also get what they want.

    All things considered, it’s one of those things that works better in theory.

    In most cases, ‘privatization’ does not really mean that a function has been given back to the market. It means that we have a highly subsidized, regulated, sometimes monopolized activity in which there is private ownership but a high degree of public control.

    Yeah, seems to fool some of the people all of the time and all of the people some of the time… I’m told that’s about as good as it get.

  8. Patrick, Alan
    Re the US, (I am not sure how/if this fits in the conversation) it is true that the congress et al, have for at least a decade, been unable to even start to consider the sort of changes to copyright law and fair use issues that are to some degree, needed because of the web and the like.

    Hence the repeated attempts to try and use their courts system to do what is really the job of elected representatives….. hasn’t worked that well, the Google Books case dragged on for years and years and years and ended in a whimper (way too ambitious )… It has now entered in to a ‘son of’ type case.

Leave a Reply

Your email address will not be published. Required fields are marked *

Notify me of followup comments via e-mail. You can also subscribe without commenting.