Small issue, little interest, move along

Last week I turned up (by video) at the Economic Development, Science and Innovation Select Committee to have my say about the Commerce Amendment Bill: supporting reform of s36 (abuse of market power), suggesting that the Commerce Commission should be able to issue quick-fix provisional authorisations like the ACCC can, and arguing against the information sharing powers being proposed for the Commerce Commission.

Very few people seem to care about the information sharing issue. There were 29 submission on the bill but on my count only 5 of them raised the information sharing powers. 

They were the Commission (for); Edward Willis of the University of Auckland Law School ("tentative support, subject to some reservations"); Malcolm Harbrow (Twitter's Idiot/Savant and No Right Turn blogger), who wanted to make sure the new provisions weren't used to subvert application of the Official Information Act (good idea); and Russell McVeagh and me (both opposed). Russell McVeagh in their executive summary say "Such expansion of powers will disincentivise businesses from voluntarily providing information to the Commission to the detriment of the current levels of efficiency in the discharge of the Commission's significant number of functions", and I agree, though I also have wider issues, as I said in my submission.

So: not many people care. And maybe it really is no biggie. At the Committee hearing the Commission sang quite a good "we're all public sector entities trying to work together" song, and I can imagine that hitting the spot. And to be fair it has guidelines (the relevant bits are pp33-4) which set a reasonably high threshold for sharing information. 

The guidelines say that "Other than as provided for in the specific situations discussed below we do not share evidence with other New Zealand or overseas law enforcement or government departments or entities". The specific situations are: joint investigations; Fair Trading Act sharing with the FMA and Takeovers Panel (allowed for in that Act); where there might be serious fraud (allowed for under the SFO's Act); where there might be serious criminal offending (sharing with the likes of the police); where there might be a serious threat to public health or public safety (whatever agency is relevant); and assisting overseas regulators the Commission has a cooperation agreement with (allowed under various Acts including the Commerce Act).

That's all sensible and desirable, and if the Commission carries on along those lines, and you'd expect it to, all good. But the problem is that the proposed legislation doesn't replicate those high standards. It just says (proposed s99AA) that 

(1) The Commission may provide to a public service agency, a statutory entity, or the Reserve Bank of New Zealand any information, or a copy of any document, that the Commission—

(a) holds in relation to the performance or exercise of the Commission’s functions, powers, or duties under this Act or any other legislation; and

(b) considers may assist the public service agency, statutory entity, or Reserve Bank in the performance or exercise of its functions, powers, or duties under this Act or any other legislation

That opens a very large door: "Any information ... that ... may assist" a public sector entity is a far looser criterion than the serious criminal offending, or serious threat to public safety, that the Commission under its present guidelines would need to see before it was prepared to share. 

Ed Willis is his excellent submission (more judicious than my own coathanger tackle) said, and I'm with him, that 

it would usually be expected that processes for the protection of relevant interests and the promotion of transparency would be incorporated into the statutory drafting itself rather than being left to the Commission to work out in practice. That said, the real issue here is one of appropriate balancing. The Committee may wish to question the Commerce Commission directly about the anticipated procedures it will employ to ensure any information sharing under the new provisions is appropriate and fit-for-purpose. Only if the Committee is satisfied with the Commission’s response should it recommend that the Bill be enacted in its current form (p5)

A practice run

We've got the Budget coming up on Thursday, so as a limbering up exercise let's have a look at how to figure out whether the Budget will be braking or boosting the economy. 

You'd think that would be one of the top things the media would cover, but in practice they don't, or not well. Far more attention goes on "the" deficit or "the" surplus, on the debt profile and fiscal sustainability over the longer haul, and on specific initiatives. They're all fine, and worth the attention, but they don't tell you whether the Budget will be helping the economy when it needs help, or tapping the brakes when it's threatening to get boisterous - the 'stabilisation' role for fiscal policy, as they say.

So as a practice run at figuring out what cyclical impact Budgets have, here's a graph of what happened in the Aussie Budget that their Treasurer Josh Frydenberg presented last week (taken from their big Budget document, Budget Paper No. 1: Budget Strategy and Outlook).

Start with the solid black line. That's the historic and forecast path of "the" deficit or "the" surplus, the headline figure commentators tend to obsess on. Indeed in both Australia and New Zealand it's become almost a macho indicator of "I can run the Treasury better than that other lot", and any Treasurer that can't point to an actual or forecast surplus must be doing it wrong. Infantile, but that's the politics of it for you.

Now have a look at the 2018-19 year, which I've helpfully identified with that green arrow.

You'll see that the headline outcome was effectively a balanced Budget, with no material surplus or deficit. But what actually happened was that the headline outcome was flattered by a stronger than usual Aussie economy. You can make a decent stab at calculating how much extra tax revenue (or how much reduced government spending on benefits) was down to the good times, and that's shown in the blue bar. It was about 1% of GDP. Without that temporary windfall, the reported headline balance turned into an underlying cyclically-adjusted deficit of about 1% of GDP (the red bar).

No great dramas either way, other than (a) illustrating the point that a Treasurer's "success" in putting a balanced or surplus Budget in the window may have nothing to do with their responsible fiscal management and rather more to do with the state of the economy and (b) reminding us all to watch what's happening to the cyclically adjusted balance as a better guide to what's really going on with the stance of fiscal policy.

Onwards to 2020-21 (the orange arrow). There's a headline deficit of a bit less than 8% of GDP, flattered a bit by cyclical conditions (including the Aussie Treasury clipping the ticket on record high iron ore prices), so the underlying deficit is actually a bit bigger, at a little over 8% of GDP.

The Aussies have quite helpfully split out the underlying deficit into two bits - how much is down to temporary anti-Covid measures (the teal-coloured bar), and how much is genuinely 'structural' (the red bar) and will still be there when things like their JobKeeper wage subsidy scheme have dropped away.

Them's the concepts. What do they tell us about the cyclical role of Aussie fiscal policy?

First of all, that in the current 2020-21 there will have been a stonking great 7% or so of GDP as anti-Covid fiscal support, on top of some 2% of GDP in the 2019-20 year. This is fiscal policy doing its stabilisation stuff, and then some. I have no problem with that: it's possible that Finance Ministers pretty much everywhere ended up overdoing it a bit, but that's okay. Providing too much counter-cyclical support is a hell of a sight better mistake to make than providing too little. 

It also shows, by the way, that fiscal policy when it puts its mind to it, can be deployed very forcefully and very quickly. That wasn't the accepted wisdom pre-Covid, but is now. 

Next thing you'll notice is that Freydenberg was in no hurry to take the Covid support away: it tails off gradually. That's probably right, too: some sectors dependent on an open border will still be hurting for quite a while yet (the Aussies' assumption in this Budget was their border not opening till mid '22). Frydenberg explicitly chose to put a higher priority on near-term cyclical support over the pace of eventual deficit reduction, and one of the key things I'll be looking for in our Budget was what choice Grant Robertson has made.

You can measure how quickly support is being taken away: it's a thing called the 'fiscal impulse', and it is the change from one year to the next in the size of the structural balance. Think of it this way: suppose this year you get $100K in wage subsidies, but next year you only get $50K because the package gets tightened up. You're still being supported, but less so. The fiscal impulse picks up that degree of tightening. For some reason the Aussies don't bother calculating it (search their Budget Paper No.1 for 'impulse' and you get nada), but we do. It's the very first figure I'll be looking for on Thursday.

Even after all the Covid support is gone, Frydenberg has also decided to take his own good time to deal to the non-Covid structural deficit, which will still be running at 1.5% - 2.5% of GDP in the later years of this decade. That's more debatable, but to the extent that an ongoing deficit reflects using once-in-a-lifetime low borrowing costs to provide infrastructure, I can live with that too. Again it'll be interesting to see what choice we make.

Oops

Every year the American Economic Association picks the year's best paper from each of its four journals (Applied Economics, Economic Policy, Macroeconomics, Microeconomics): there's a list of the winners for the past decade here if you'd like to catch up with the good stuff. In a nice touch, you can download the full text of each one without being an AEA member. 

So us pro-competition types can feast, for example, on the finding by the 2016 Economic Policy winner ('Death by Market Power: Reform, Competition, and Patient Outcomes in the National Health Service') that increased competition in the UK's National Health Service worked:

Within two years of implementation, the NHS reforms resulted in significant improvements in mortality and reductions in length of stay without changes in total expenditure or increases in expenditure per patient. Our back of the envelope estimates suggest that the immediate net benefit of this policy is around $479 million per year

Fernando Luco, who hails from Texas A&M University, is this year's winner in the Microeconomics category with 'Who Benefits from Information Disclosure? The Case of Retail Gasoline'. This one, though, doesn't have such an unambiguously happy moral.

Our story starts in Chile in 2012, when the government required petrol stations to post their prices on a government website and update them promptly (within 15 minutes) whenever they changed. An everyday tale of empowering customers to find the best deal, you might think.

Except that petrol stations' margins increased by some 9% after the new disclosure policy, and various statistical checks confirmed that it was indeed the disclosure policy, and not some non-policy event happening at the same time, that was behind the petrol stations coining it.

What went on? As the author says (p278)

information disclosure may have both pro- and anti-competitive effects. On the one hand, disclosure may intensify competition if consumers benefit from lower search costs and firms use the website to compete more intensively. On the other hand, if stations can easily monitor their rivals’ actions and consumers do not actively use the disclosed information, disclosure may facilitate coordination

so it could go either way, and the net effect depends on who uses the info smartest.  As it happens, the author had a dataset of smartphone price look-ups, geocoded, so he could see in each local petrol market how actively people were checking out the prices on offer. In aggregate, the petrol stations won - but not everywhere (p302):

price disclosure allowed firms to monitor their rivals’ actions and to increase their payoffs on average. However, when consumers actively engaged in search [which he could tell, from his dataset], the demand-side response to disclosure dominated and competition intensified

The more smartphone-savvy higher income areas came out okay, lower-income areas not so much: "while in the lowest income areas margins increased by around 12 percent, margins decreased by 4 percent in the highest income areas, suggesting that disclosure affected low income areas the most" (p296). 

Bottom line from a competition policy point of view (p303)? 

mechanisms that increase market transparency may increase competition and benefit consumers only if consumers can easily access and use the disclosed information. Otherwise, the supply-side response to disclosure is likely to dominate, and the intensity of competition will decrease. Hence, this paper provides evidence showing that policy makers should consider ease of access to the newly disclosed information to be of major importance

In a New Zealand context, I suspect smartphone use probably doesn't vary as much with income as (I'm guessing) it does in Chile. Even so, the general proposition still applies: a price disclosure Cunning Plan has the potential to be a good pro-competition pro-consumer idea, but only if there's a good deal of effort put into making sure people from every walk of life end up using the thing.

The pointy end

Almost seven years ago,  the Productivity Commission raised it as an issue worth looking at. MBIE put out an issues paper in 2015; submissions rolled in in early 2016; the government kicked for touch in 2017; MBIE put out another discussion paper in 2019; a policy paper went to Cabinet in February 2020; and at long last a Bill was introduced this month. The Commerce Amendment Bill 2021 is now with the Economic Development, Science and Innovation Select Committee, which is due to report back by September 16. 

In sum, we've finally got to the pointy end of doing something about s36, the bit of our Commerce Act that is supposed to rein in powerful incumbents from abusing their market power, but doesn't. The Bill provides for changing our s36 wording from the current unsatisfactory

"A person that has a substantial degree of power in a market must not take advantage of that power for the purpose of — (a) restricting the entry of a person into that or any other market; or (b) preventing or deterring a person from engaging in competitive conduct in that or any other market; or (c) eliminating a person from that or any other market"

to

"A person that has a substantial degree of power in a market must not engage in conduct that has the purpose, or has or is likely to have the effect, of substantially lessening competition in — (a) that market; or (b) any other market".

It gets us out from under all the problems "taking advantage of" has caused in our competition jurisprudence, and for good measure lines us up with the Aussies, who made the same change to their equivalent legislation back in late 2017.

Submissions to the Select Committee are due by April 30, so it's time to start putting your thoughts together. If you want to keep track of how the Bill is going, you can sign up to get e-mail alerts from the Select Committee here and you can follow progress at the Bill website.

I will be submitting in wholehearted support of the change to s36 for the reasons I mentioned last year when the policy paper went to Cabinet ('It creeps ever closer'), and I hope others will, too. But I'll be opposing one element of the Bill. Last year I didn't like the look of a proposed amendment, giving the Commerce Commission powers to share information it holds with other public agencies: I thought that "there should be a tough threshold test before any of that information gets passed around the wider public sector agencies", and indeed the Cabinet paper had talked about "appropriate safeguards".

In fact the proposed level of safeguarding is pitiful. The Bill at s99AA(1)(b) provides that the Commission can hand information over when it considers it "may assist the public service agency, statutory entity, or Reserve Bank in the performance or exercise of its functions, powers, or duties under this Act or any other legislation".

"Hey, guys, you might find this handy" is no safeguard at all.

What's next?

The OECD came out this week with an update to its economic outlook and a set of policy recommendations. If you've wondered how New Zealand has been faring through the Covid outbreak, relative to my selection of the usual suspects, here's how we went in 2020 (we're red, and there are a couple of comparators in green, the OECD, and the world as a whole) ...

and here's how the OECD sees 2021 going ...

... and here's their best guess at 2022.

We scrub up pretty well when you look back at 2020: we had a typical-sized downturn (a tad worse than the world as a whole, a tad better than the typical higher-income OECD country), but for a much better Covid outcome than in most places, so as a package it's pretty impressive. Longer run, it looks, unfortunately, like back to the pre-Covid status quo: the OECD says we should expect relatively slow growth by either world standards or by comparison with our better-off peers.

On policy, beyond the obvious big macro settings (" A premature tightening of fiscal policy must be avoided.  The current very accommodative monetary policy stance should be maintained"), the OECD's recommendations are

• "vaccinate fast" (Co-ordinate and accelerate vaccination of adults across the world, ensure poor countries receive their fair share of doses, and improve funding for the COVAX initiative, ensure effective test, track and trace programmes)

• "invest fast" (Speed up implementation of new spending plans...to boost growth and jobs, help businesses adapt to a digital future, privilege grants and equity-type support over debt to give viable small and medium-sized companies the space to develop, invest in cleaner infrastructure and digital technology to foster a transition to a more resilient and sustainable economy) and

• "support people" (Protect the incomes of people hit hardest by the crisis, help the low skilled and the vulnerable, improve training schemes and access to the labour market, focus on youth – young people need particular support now and to help them prepare for a changing world of work)

and who's going to argue with any of that, though you have to feel that "invest" and "fast" do not come easy to New Zealand policymakers, and I'll be pleasantly surprised if it happens. It would be nice if, for once, we just took the OECD's ideas and ran with them quickly and comprehensively: unfortunately, our track record is not crash hot ('Are we serious?', 'Take advice? Moi?').  Among other things, we wouldn't have today's housing market problems if we'd listened to the OECD's suggestions, going back to 2017, on how to address them. 

What if they threw a party ...

... and nobody came?

Which, in the competition space, is where we've got to, with our attempt to make it easier for companies to collaborate to deal with problems like Covid disruption to supply chains. We set things up to make it easier and faster for the Commerce Commission to authorise collaboration - but nobody's used the new dispensation.

Here's the background. The COVID-19 Response (Further Management Measures) Legislation Act 2020 went live mid May of last year. Its heart was in the right place. During an 'epidemic period', a new s65AD of the Commerce Act allows the Commission to issue a provisional authorisation. ss65AD(2) and (3) are the key bits:

(2) The Commission may make a determination in writing granting a provisional authorisation ... if the Commission considers it is appropriate to do so—

(a) for the purpose of enabling due consideration to be given to the application; or

(b) for any other reason.

(3) The Commission is not required to comply with section 61(5) to (6A) before granting a provisional authorisation

(3) means that the Commission doesn't have to hang about and be "satisfied" - the test in normal times, under s61(6) - that the collaboration would provide "a benefit to the public which would outweigh the lessening in competition that would result". It can just say, get on with it for now and we'll do a fuller analysis later. Right on.

As fellow aficionados of New Zealand's mania for micromanagement will appreciate, you don't see the wide discretion of "for any other reason" written into the Kiwi statute books too often. In fact it's a straight crib from the Aussie equivalent (for wonks, s91(2)c of their Competition and Consumer Act). The ACCC has been able to do "interim" authorisations all along, Covid or no Covid, and you'd think that's a sensible plan.

Oddly, our own Commission used to have the power to issue provisional authorisations under the then s63 of the Commerce Act, but that got repealed by s22 of the Commerce Amendment Act 1990. Go figure.

Never mind, here we are today, back to the original status quo, even if it only applies in epidemic periods. Given that in normal times getting an authorisation strongly resembles wading through hip-high tar for months carrying a complete bound set of Econometrica, something cheap, cheerful and, above all, fast, was just what was needed to enable urgent collaboration in the public interest.

But nobody's used the new process. There have been no applications for authorisation under the new provisions.

Which is odd, because over in Oz, the ACCC has been issuing interim authorisations all over the place: public and private hospitals, medical wholesalers, grocery retailers, banks, regional airlines, and fuel importers and distributors. It's been commendably quick - overnight, on one occasion - in turning the applications around, and it's been putting pro-competitive safeguards such as time limits into them without nobbling the public welfare point of it all. 

What sort of stuff is being authorised? For the supermarkets, as an example, it was things like jointly addressing panic buying: "co-ordinating store hours, including allocating dedicated shopping hours for  elderly and disadvantaged members of the public during periods of high demand for Retail Products", "implementing uniform or similar purchase limits and related public messaging", and "measures to ensure continuity of supply to consumers in remote or regional areas, including securing special allocations of stock and joint requests to suppliers", as you can read in this draft determination.

Sure, Australia and New Zealand each have their own funny little ways, and there's no reason why we should do everything the same way. But I'm still bemused why helpful interim authorisations are flying off the ACCC shelves, while nobody's come into the Commerce Commission looking for one. Anyone got ideas?

An OCR cut isn't dead after all

Every economist polled before today's Monetary Policy Statement (eg here and here) said there'd be no change to the Official Cash Rate (OCR). Quite a few (including me) thought it's unlikely that any further cuts are on the table (and that the option of going to a negative OCR might even be taken off the table by the Reserve Bank); that the next move will be up (likely preceded by winding back some of the unconventional monetary policy tools); but that it's still necessarily a longish way away. Personally I was also wondering what weight the Bank would place on the December higher than expected inflation rate and the very much lower than expected unemployment rate.

The OCR did indeed stay at 0.25%. Taking it to zero or into negative territory has not, however, taken off the table: quite the contrary. The Statement said (p3) that it's now operationally do-able ("the operational work to enable the OCR to be taken negative if required is now completed") and the minutes of the latest policy meeting said (p6) that "The Committee agreed that it was prepared to lower the OCR to provide additional stimulus if required". There was even a section on pp26-7 headed "A zero or negative OCR would provide additional monetary stimulus, if required" and went into how it might work out in practice.

Adrian Orr was asked at the media conference about the pre-MPS expectation that a lower OCR might be sidelined: he said the bank prefers to maintain a full suite of options against whatever might befall down the track. Fair enough: personally I was starting to wonder whether a lower OCR catalyst might be the need to defuse any unwanted appreciation in the Kiwi dollar. 

The Bank is rightly being cautious on reading too much into the December numbers: "How long the recent recovery in inflation and employment can be sustained is highly uncertain ... Our baseline scenario for the economy, while starting from a stronger position than assumed in November, is subdued. Significant monetary stimulus remains necessary to confidently and sustainably meet our inflation and employment objectives" (p7). 

So there will be "a prolonged period of time to pass before these [inflation and employment] conditions are met" (p6) and "Meeting these requirements will necessitate considerable time and patience" (p3). How long? The Bank publishes a measure called the "unconstrained OCR"* which has a go at measuring the overall degree of stimulus: its latest one (p41, reproduced below) shows the existing level of stimulus increasing just a tad during the rest of this year, but gradually becoming less stimulatory (though still strongly supportive) through 2022 and into 2023. When will monetary policy be back at "neutral"? If (as Figure 6.4 on p50 suggests), a "neutral" OCR is something like 2%, chances are that we won't see one till 2024 or beyond.

There's usually some interesting one-off stuff in each Statement. There's a box (pp19-21) on current conditions in the Māori economy, and this new visualisation (p14) of conditions in the labour market, which does a nice job of bringing together the various ways you might look at the market to figure out how close you are to maximum sustainable employment. 

*In the May 2020 Statement, p11, it had said that "the Reserve Bank used a projection of the OCR to highlight the level of monetary stimulus needed to achieve our inflation and employment objectives. A fall in the OCR projection relative to the previous Statement meant that more policy stimulus was needed. We have had to modify this practice given the Monetary Policy Committee’s forward guidance on the OCR out to early next year and the use of alternative monetary policy instruments. We have opted to publish an unconstrained OCR ... This demonstrates the broad level stimulus needed to achieve the Reserve Bank’s monetary policy objectives, much like the OCR projection demonstrated in the past"

You probably got ripped off, too

I'm involved with a training programme for an overseas competition agency, and we've been going through the law and economics of cartels: I've been talking about the detriments they cause. As it happens, there's a recent spectacular example of how bad they can be.

One of the gurus of cartel documentation has been John Connor, emeritus professor at Purdue and a senior fellow of the American Antitrust Institute. In 'Twilight of Prosecutions of the Global Auto-Parts Cartels', he's written about one of the biggest cartels ever (amounts cited are US dollars):

At last count 18 jurisdictions vigorously prosecuted this supercartel, which demonstrated exceptional duration, global reach, size, and injuriousness. Estimates for affected commerce of the Auto-Parts supercartel range from $3.2 to $5.0 trillion. There are few reliable estimates of overcharges, but averaging the few preliminary estimates suggests that injuries are in the range of $0.6 to $1 trillion (p1)

That's a pretty impressive level of overcharges, exacted from 17 different car manufacturers from around the turn of the millennium though to late 2009, when the cartels started to get rumbled. Chances are, if you bought a car in the 2000s, you're one of the vast number of victims yourself. Depending on what criterion you use, this was either the largest or second largest grouping of related cartels ever found.

Connor reproduced this picture, originally from a European Union enforcement announcement, showing just how many different components of a car got targeted by the cartels. Basically everything but the engine, the transmission and the bodywork got rorted. If the 'wire harnesses' rings a bell, that's one where the ACCC got involved: the Japanese company pinged in that case unwisely decided to appeal their fine, with appropriately karmic results.

There's a great deal of stuff to take away from this example, one being that companies can be too hard-nosed at bargaining for their own good: "Did the assemblers [the carmakers] push too hard on price reductions in the early 2000s, and thereby trigger supplier collusion to cope with an existential threat?" (p1). It's not a defence, and it doesn't even take us very far into tout comprendre c'est tout pardonner territory, and you shouldn't blame the victims, but (like the Lodge case here at home) you do wonder whether the damn thing would ever have happened in a more reasonable world.

And if you're into corporate strategy, you might want to think twice about all that financial engineering cleverness. You can't be rorted if you've vertically integrated that component. Divest it, though  - free up capital and all that good stuff - and you've opened up some vulnerabilities:

Perhaps it is no coincidence that collusion against GM began soon after it sold its Delco-Remy division in 1994, against Toyota after DENSO no longer supplied a majority of output to Toyota, etc. Any financial cost savings resulting from divestment may have backfired. Second, by delegating manufacturing of minor inputs (say, 2% or 3% of total material inputs) to ostensibly unaffiliated suppliers, the OEMs ['original equipment manufacturers', i.e. the carmakers] suffer a substantial loss of information about manufacturing costs and a consequent loss of bargaining power over price (p16)

The main takeaway from the cartel, though, is the ineffectiveness of the fines. We know the theory: if there's a 100% chance of being detected, a fine equal to 100% (or more) of the overcharging will deter; if there's a 20% probability, then a fine of 500% (or more). What actually happened here?

There are few reliable estimates of the size of overcharges for these cartels, and more are desperately needed, but averaging the few preliminary estimates suggests that injuries are in the range of $0.6 to $1 trillion. Even if monetary penalties rise to double the current $20 billion, cartel deterrence or cartel dissuasion is highly unlikely (p30)

Which does get you thinking about the alternative sanction of criminalisation - sending executives to jail - which we are about to introduce from April. In egregious cases like car parts, you can see why people would well reach for it. But as someone from the socially liberal end of town, I have two reservations.

Here's the first one (data from this Statista page, and originally compiled here). We're already well down the wrong end of this international comparison: we're just about the last country in the world that ought to be thinking about putting even more people in jail. 

And here's the other:

Unlike typical international cartel prosecutions in the past, few of the individuals held accountable by antitrust authorities in Auto Parts have been CEOs, COOs, or CFOs of their parent companies or even their subsidiaries ... Rather, they have held titles like sales manager, director, marketing manager, or department head of units below the corporate VP level (p27).

When we eventually press the criminalisation button, I hope we don't make the car parts mistake - jailing some midlevel functionaries, but letting the big fish swim free.