8 ms·

The IFF referenced in the article seems like a great example of how good-intentioned changes to incentive structures can have very warped outcomes, potentially

The IFF referenced in the article seems like a great example of how good-intentioned changes to incentive structures can have very warped outcomes, potentially years later. These kinds of effects keep popping up for me, even in industry contexts (e.g. stack ranking). Are there accepted mechanisms for systematically identifying these knock-on effects, and if so, what are they and how can they be more broadly applied? How many "hops" of influence can you get away from the change before the effects are impossible to predict? Or does it just boil down to "ask very smart domain experts to think about the problem very hard for as long as you can afford to pay them"?

iudqnolq 7y ago
The key quote this is responding to: > ... less [to do] with McKinsey and more with an incentive to overpay more generally. It’s more likely something called the ‘Industrial Funding Fee,’ or IFF. The GSA’s Federal Acquisition Service gets a cut of whatever certain contractors spend using the GSA’s schedule, and this cut is the IFF. The IFF is priced at .75% of the total amount of a government contract. In the case of McKinsey, since 2006, “FAS has realized $7.2 million in Industrial Funding Fee revenue.” > ... The IFF also incentivizes the GSA to get the government to outsource to contractors anything it can, simply to get more budget. The IFF has been creating problems like the McKinsey over-payment for a long time. In 2013, the GSA Inspector General traced a similar situation with different contractors. Managers at GSA overruled line contracting officers to raise prices taxpayer pay for contractors Carahsoft, Deloitte and Oracle. Government managers at GSA micro-managed and harassed their subordinates and damaged the careers of contracting officers trying to negotiate fair prices for the taxpayer. > How did the GSA get such a screwed up incentive...?... [T]o become more entrepreneurial as part of its “Reinventing Government” initiative, Bill Clinton’s administration implemented the Industrial Funding Fee structure. It worked in generating money for the GSA ... so well that Congress’s investigative agency found in 2002 that the GSA stopped having to rely on Congressional appropriations.
TheSoftwareGuy 7y ago
>Are there accepted mechanisms for systematically identifying these knock-on effects, and if so, what are they and how can they be more broadly applied? Yes, look for cases where A is rewarded, but B is expected: https://www.ou.edu/russell/UGcomp/Kerr.pdf https://www.ou.edu/russell/UGcomp/Kerr.pdf
drak0n1c 7y ago
That's already in practice - there are dozens of competing think tanks that churn out pretty deep critiques of legislation before it comes up for a vote. Whether for better or worse instead of the legislation being edited accordingly the criticisms are usually dismissed as "Republican/Democrat talking points".
smacktoward 7y ago
In terms of lawmaking, the usual way to deal with this problem isn't with front-end analysis, but rather by adding a "sunset provision" (see https://en.wikipedia.org/wiki/Sunset_provision https://en.wikipedia.org/wiki/Sunset_provision). That's a clause in a bill that requires it to be re-authorized periodically in order to stay in effect. If a law with a sunset provision ends up causing unintended consequences, then lawmakers can let it die just by doing nothing. That's an easier lift than you get in a bill without a sunset provision, which can only be killed if you can convince a majority of lawmakers to actively kill it. Sunset provisions aren't nearly as widely used as they probably should be.
- codesforhugs 7y ago
  If I were writing the constitution for a representative democracy today, I think it would include mandatory sunsets for all legislation. Yes, this would lead to the legislature spending a lot of time just reauthorizing existing legislation, but I would argue that the majority of new legislation being passed in modern representative democracies would be better off as revamps of existing legislation anyway.
  - johnnyo 7y ago
    This assumes you don’t have gridlock in the House/Senate, so it’s a bad idea. We already have govt shutdowns when we can’t pass a budget, imagine if a law that legalized gay marriage was allowed to expire, or one that granted people health care or immigration status. Even if a non-discrimination law only expired for a short period of time, that would be enough to do real damage to people’s lives.
    - codesforhugs 7y ago
      Gridlock problems are particular to two party systems. A modern constitution would also include voting methods that don't result in a two party system. You could also reduce the risk by the constitution itself being exempt from sunsetting and enumerating crucial rights and services therein. Anyway it's not something that's ever going to be implemented anywhere, so it doesn't really matter if it could work.
tootie 7y ago
The piece I don't get is why the agencies didn't revolt. I get how it's a warped incentive for the GSA but isn't the GSA working at the behest of agencies like ICE? Shouldn't ICE be up in arms that their budget is being squandered?
- csa 7y ago
  Oversimplified answer... Because the agencies were (and still are) told that they must use GSA first. The theory is that GSA market will provide a competitive price or that the GSA sole source will be a competitive price. Reality does not match the theory.
beaconstudios 7y ago
the general topic would be systems theory / systems thinking. Lots of little scraps of knowledge of these kinds of knock-on effects are scattered around in various fields and disciplines, but I don't know of anywhere where they're gathered in one place for broad consideration or application. Not yet, anyway.
mac01021 7y ago
I might just be dense, but I don't fully understand how the IFF was supposed to work. What was its intended effect? Is it just a tax that the contractors have to pay on spending the government's money, to disincentivize them to spend that money?
- smacktoward 7y ago
  It was intended to make the GSA self-sustaining in budget terms. Under the old system, the costs of running the GSA came out of the Federal budget. The goal back in the '90s was to reduce expenditures in order to cut down the budget deficit. IFF let them push the cost of running GSA off of the Federal budget, by pushing it onto the budgets of the vendors winning GSA contracts.
codesforhugs 7y ago
The fundamental problem with incentives is that they're asymmetric in nature: The incentivee has a lot more time (and direct motivation) to come up with a way to game the incentive than the incentivizer can spend when setting it. The only real way to address that is to revise incentives on a frequent and regular basis — but who wants to do that? Certainly not legislatures.
- wiseleo 7y ago
  That's true. I found a lucrative opportunity recently and identified and removed the bottlenecks that were throttling my earnings through bad processes on the client's side. As a software and network engineer, I found much faster ways to get things done. I am working on an app to make it even faster. That is what flat rates do - the side willing to work for a flat rate optimizes the processes to make that rate work for them. Now everything gets resolved in one service call, but my effective hourly rate is high enough to work around every inefficiency. I have an incentive to find everything that is wrong with that location and fix it regardless of what the original service call description says. That is just a starting point as far as I am concerned. If that company invested a small amount in producing technical training for field technicians and bolstering staffing at their inexpensive offshore call center, I would have had meaningful competition. I imagine if they hired consultants, they would see that as a report. I hold unique institutional knowledge like undocumented direct phone numbers to relevant people at client organizations that I have a very big incentive not to share.

iudqnolq 7y ago

The key quote this is responding to: > ... less [to do] with McKinsey and more with an incentive to overpay more generally. It’s more likely something called the ‘Industrial Funding Fee,’ or IFF. The GSA’s Federal Acquisition Service gets a cut of whatever certain contractors spend using the GSA’s schedule, and this cut is the IFF. The IFF is priced at .75% of the total amount of a government contract. In the case of McKinsey, since 2006, “FAS has realized $7.2 million in Industrial Funding Fee revenue.” > ... The IFF also incentivizes the GSA to get the government to outsource to contractors anything it can, simply to get more budget. The IFF has been creating problems like the McKinsey over-payment for a long time. In 2013, the GSA Inspector General traced a similar situation with different contractors. Managers at GSA overruled line contracting officers to raise prices taxpayer pay for contractors Carahsoft, Deloitte and Oracle. Government managers at GSA micro-managed and harassed their subordinates and damaged the careers of contracting officers trying to negotiate fair prices for the taxpayer. > How did the GSA get such a screwed up incentive...?... [T]o become more entrepreneurial as part of its “Reinventing Government” initiative, Bill Clinton’s administration implemented the Industrial Funding Fee structure. It worked in generating money for the GSA ... so well that Congress’s investigative agency found in 2002 that the GSA stopped having to rely on Congressional appropriations.

TheSoftwareGuy 7y ago

>Are there accepted mechanisms for systematically identifying these knock-on effects, and if so, what are they and how can they be more broadly applied? Yes, look for cases where A is rewarded, but B is expected: https://www.ou.edu/russell/UGcomp/Kerr.pdf https://www.ou.edu/russell/UGcomp/Kerr.pdf

drak0n1c 7y ago

That's already in practice - there are dozens of competing think tanks that churn out pretty deep critiques of legislation before it comes up for a vote. Whether for better or worse instead of the legislation being edited accordingly the criticisms are usually dismissed as "Republican/Democrat talking points".

smacktoward 7y ago

In terms of lawmaking, the usual way to deal with this problem isn't with front-end analysis, but rather by adding a "sunset provision" (see https://en.wikipedia.org/wiki/Sunset_provision https://en.wikipedia.org/wiki/Sunset_provision). That's a clause in a bill that requires it to be re-authorized periodically in order to stay in effect. If a law with a sunset provision ends up causing unintended consequences, then lawmakers can let it die just by doing nothing. That's an easier lift than you get in a bill without a sunset provision, which can only be killed if you can convince a majority of lawmakers to actively kill it. Sunset provisions aren't nearly as widely used as they probably should be.

codesforhugs 7y ago
If I were writing the constitution for a representative democracy today, I think it would include mandatory sunsets for all legislation. Yes, this would lead to the legislature spending a lot of time just reauthorizing existing legislation, but I would argue that the majority of new legislation being passed in modern representative democracies would be better off as revamps of existing legislation anyway.
- johnnyo 7y ago
  This assumes you don’t have gridlock in the House/Senate, so it’s a bad idea. We already have govt shutdowns when we can’t pass a budget, imagine if a law that legalized gay marriage was allowed to expire, or one that granted people health care or immigration status. Even if a non-discrimination law only expired for a short period of time, that would be enough to do real damage to people’s lives.
  - codesforhugs 7y ago
    Gridlock problems are particular to two party systems. A modern constitution would also include voting methods that don't result in a two party system. You could also reduce the risk by the constitution itself being exempt from sunsetting and enumerating crucial rights and services therein. Anyway it's not something that's ever going to be implemented anywhere, so it doesn't really matter if it could work.

tootie 7y ago

The piece I don't get is why the agencies didn't revolt. I get how it's a warped incentive for the GSA but isn't the GSA working at the behest of agencies like ICE? Shouldn't ICE be up in arms that their budget is being squandered?

csa 7y ago
Oversimplified answer... Because the agencies were (and still are) told that they must use GSA first. The theory is that GSA market will provide a competitive price or that the GSA sole source will be a competitive price. Reality does not match the theory.

beaconstudios 7y ago

the general topic would be systems theory / systems thinking. Lots of little scraps of knowledge of these kinds of knock-on effects are scattered around in various fields and disciplines, but I don't know of anywhere where they're gathered in one place for broad consideration or application. Not yet, anyway.

mac01021 7y ago

I might just be dense, but I don't fully understand how the IFF was supposed to work. What was its intended effect? Is it just a tax that the contractors have to pay on spending the government's money, to disincentivize them to spend that money?

smacktoward 7y ago
It was intended to make the GSA self-sustaining in budget terms. Under the old system, the costs of running the GSA came out of the Federal budget. The goal back in the '90s was to reduce expenditures in order to cut down the budget deficit. IFF let them push the cost of running GSA off of the Federal budget, by pushing it onto the budgets of the vendors winning GSA contracts.

codesforhugs 7y ago

The fundamental problem with incentives is that they're asymmetric in nature: The incentivee has a lot more time (and direct motivation) to come up with a way to game the incentive than the incentivizer can spend when setting it. The only real way to address that is to revise incentives on a frequent and regular basis — but who wants to do that? Certainly not legislatures.

wiseleo 7y ago
That's true. I found a lucrative opportunity recently and identified and removed the bottlenecks that were throttling my earnings through bad processes on the client's side. As a software and network engineer, I found much faster ways to get things done. I am working on an app to make it even faster. That is what flat rates do - the side willing to work for a flat rate optimizes the processes to make that rate work for them. Now everything gets resolved in one service call, but my effective hourly rate is high enough to work around every inefficiency. I have an incentive to find everything that is wrong with that location and fix it regardless of what the original service call description says. That is just a starting point as far as I am concerned. If that company invested a small amount in producing technical training for field technicians and bolstering staffing at their inexpensive offshore call center, I would have had meaningful competition. I imagine if they hired consultants, they would see that as a report. I hold unique institutional knowledge like undocumented direct phone numbers to relevant people at client organizations that I have a very big incentive not to share.