Pre-qualification for Bandar Lampung Bulk Water Supply project is open

I'm not sure how I missed this (or how the English, and Indonesian language media also seems to have entirely missed it at time of writing), but the Bandar Lampung Bulk Water Supply public private partnership ("PPP") project pre-qualification has been open since 30 March, and is open until 30 May 2017.

Source: PDAM Way Rilau

It's odd that this isn't bigger news because, as far as I am aware, this is the first PPP project to be tendered since the Palapa Ring projects were tendered in late 2015

This is the second time this project has been tendered, and the previous go-around was quite a big deal.

What happened last time

The previous tender process was launched in 2012, and attracted substantial interest from domestic and foreign bidders, with 10 consortia submitting bids. Shortlisting happened fairly quickly, with Water Consortium (Hyundai Engineering and Construction, Itochu Corporation and PT Potum Mundi Infranusantara); Abeima and PT Wijaya Karya Persero Tbk; Acuatico and Mitsubishi Corporation; and Manila Water and Great Giant Pineaple Co forming the final four.

Yet, disagreements over the project costs; the allocation of those costs between the users, central government, and city government; and the project structure meant that the final request for proposal was substantially delayed.

The issue of the project costs was arguably the most critical issue, with widely diverging capital costs proposed by different parties within government. The Ministry of Finance allocated a maximum subsidy based on the lower estimate of the capital costs--allocating a maximum subsidy of IDR 350 billion--and the project went to tender in September 2015.

The Ministry of Finance's estimate of the required subsidy proved to be excessively conservative, and none of the bidders ended up submitting conforming bids, resulting in a failed tender.

Why did no one submit a conforming bid?

Indonesia's main way to provide subsidies to infrastructure projects is through what they call Viability Gap Funding ("VGF"). The aim of this VGF is to bridge the financial "viability gap", allowing government to engage the private sector to deliver a project that they otherwise wouldn't be interested in. (Note: there's a lot more to say about VGF, and why, and when it should be offered, but that's a for separate blog post sometime.)

Indonesia's VGF guidelines are interpreted to mean that VGF is not allowed to exceed a maximum of 49% of the capital cost of a project. This means we that, based on the announced IDR 350 billion VGF, the government was assuming the capital costs for the project were IDR 700 billion (with a little rounding).

The graph below illustrates how the tender failed.

In this project, the revenues from PDAM are fixed*, the bid factor becomes the minimum VGF required to deliver a project that can fulfill the technical specifications. Whichever bidder gets the optimisation of the lowest capital, operating, and finance costs, should be able to offer the lowest required VGF to do the project, and win.

The Owner’s Estimate column showing project costs of IDR 1.4 trillion is present value terms, and equivalent revenues shows a simplified version of the project cashflows that illustrates how the government arrived at the IDR 350 billion VGF number. For this simple illustration, I’m assuming that the total operating expenditure over the life of the project is the same as the capital expenditure. If someone feels like giving me a copy of the government’s owner’s estimate financial model, I’ll happily update the figures, but it doesn’t really matter for the illustration.

The Government’s Ideal Scenario shows what the government was hoping for, where some party ideally was able to get the project done for less than they estimated, and would bid with a VGF of less than IDR 350 billion.

The Actual Lowest Bid shows what actually happened. The bidders would have been working their hardest to get the lowest capital, operating and finance costs to decrease the VGF required to meet their costs, and get under the government’s cap, but none of the four was able to do it. So, the tender failed.

Failed tenders are really expensive

The failure of the Bandar Lampung Bulk Water Supply Project destroyed a lot value in the Indonesian economy; for the government, for the bidders, but--most importantly--for the citizens of Bandar Lampung.

On the government's side, they had multiple rounds of advisers, including top-shelf (and quite expensive) commercial, legal, and technical advisers, both for the preparation of the transaction, and for the appraisal of that transaction. In aggregate, the fees paid to these advisers would be well into the millions of dollars, and I'd be surprised if it were below USD 10 million. Beyond the fees, there is the wasted time of all of the civil servants, all the way up to the minister level, in multiple ministries and agencies, who had to participate in meetings, and produce their own analysis and recommendations.

For the bidders, participating in tender processes is also expensive. Preparing a pre-qualification submission can easily run into the multiple hundreds of thousands of dollars each in staff time and advisers' fees to make sure they can do the project, and to prepare the documents in the format required. Then, the shortlisted parties easily blow another couple of million dollars to prepare a conforming bid, including their preliminary designs. With 10 pre-qualification submissions, an extended pre-bid process, then four short-listed bidders, there would easily be another USD 10 million dollars of value wasted there.

Finally, and most importantly, the citizens of Bandar Lampung have had to go another five years without adequate water supply as a result of these delays, wasting money buying water from eceran, wasting time boiling water and otherwise creating plans to make sure they can access potable water as and when they need it, and bearing health and lost productivity costs due to the poor quality water.

Source: Statistik Kesejahteraan Rakyat Kota Bandar Lampung 2016, table 6.6, page 100

As it is, only around 20% of Bandar Lampung's 1.2 million citizens have access to piped water for cooking, or washing, and pretty much all other potential sources are worse than PDAM supply by some metric; either more expensive (private wells with jetpumps), take a lot more time (wells), or are much lower quality (unprotected wells). Although, on the other hand, most of them are more reliable. 

The estimation of economic benefits of infrastructure projects is always hazy, with lots of uncertainties, but we can come up with a ballpark figure using some simple arithmetic.

The original project was for a 500 litres-per-second facility, aimed at serving 44,000 households. This is far less than the 960,000 people currently without service in Bandar Lampung, so we can assume 100% take-up pretty quickly. So, 44,000 households went with poor quality water for 5 years. If we assume the costs of the poor quality water are a dollar a day per household (counting all the wasted time, purchasing of bottled water, health impacts, etc. etc. that's certainly too low), that's already USD 80 million.

So, just as a conservative ballpark estimate of the cost of the delay, we can get to USD 100 million pretty quickly. Suffice to say that the costs of inaction are really high, and orders of magnitude higher than a couple of millions of dollars of VGF.

What has changed this time?

The word on the street is that the project has been rescoped, and redesigned to make it deliverable under Indonesia's current PPP framework. Just from the details available in the Request for Pre-qualification document on the PDAM Way Rilau website, we can see that there have been some changes made.

The raw water intake is now 825 litres-per-second, and the water treatment plant is 750 litres-per-second, up from the 500 litres-per-second of the last go around. The extra volumes mean that there are potentially more revenues to go around, to decrease the required VGF.

However, the capital costs of the project still seem to be estimated at IDR 700 billion. I'm not entirely sure how the plant capacity increases, and the capital costs (that were already proven to be too conservative) stay the same, but I'm no engineer, so what do I know? It's possible that they have found efficiencies in other areas, or are trading off capex for opex.

If you want in, you'd better hurry up

The pre-qualification process is open for another two weeks from today, so if you are planning on bidding, you'd better get on it!

---

* Technically not correct, but effectively the same. There was a fixed payment to cover the capital costs, and a variable payment to cover the variable costs of operation. This aims to ensure there is no economic profit (or loss) coming from the operation of the facility if the contractor performs satisfactorily.

The Indonesian government just needlessly cost itself and its citizens a hundred million dollars

The Indonesian government announced last week that they will be cutting toll road tariffs by 25% on private toll roads, and 35% on Jasa Marga run toll roads, effective from the 7th of July until 5 days after Lebaran.

The government instructed toll road operators to make the cut to help out people during their long trip home.

In this post, I’m going to argue that, yes, this does save some of Indonesia's citizens money, but that it is inefficient, poorly targeted, and will cost the government (at least) a hundred million dollars over the next few years.

First off, how much does it save people?

I don’t have numbers on the amount of money this could cost toll road operators (which is the same as the aggregate amount that it saves users), but we can try to estimate a ballpark figure.

The tariff cut is in effect for 16 days, being 16/365 = 4.4% of a year. How much money the companies will lose will depend on the traffic flows. If traffic flows are heavier over that period than over the year as a whole, we’ll lose 25%/35% on more than 4.4% of a year, if they’re light then they will lose less than that.

To me, it seems plausible that they’d be heavy, because a large chunk of the country will be travelling over this period. But then, most toll roads are in Jakarta, and people leave the city during this period, and the flows might be heavy only for a few travelling days, while they’ll be light for a few days in the middle; this could mean that the flows are on average lighter. For this sort of analysis, it’s probably not too far wrong to just assume that traffic flows are, over the whole 16 day period, pretty much in line with the average flows over the year. In that case, we can just multiply the discount by 4.4% of the year, and find that private toll road operators should expect to lose around about 1.1% of their annual revenues over this period, while Jasa Marga will lose 1.5%.

I haven’t got data on the total fare revenue collections on all toll roads in Indonesia, but, in 2014, Jasa Marga’s receipts were around IDR 6.6 trillion and they control a bit under 60% of Indonesia’s toll road network. Assuming revenues roughly split that way too, we can estimate around IDR 4.8 trillion of non-Jasa Marga toll revenues. With a bit of growth and rounding, I’ll assume IDR 7 trillion for Jasa Marga and IDR 5 trillion for non-Jasa Marga toll operators.

So, as a result of this tariff cut, Jasa Marga will lose around IDR 105 billion (USD 8 million), and private companies will lose around IDR 55 billion (USD 4 million) in revenues. Toll road users will be better off by around IDR 160 billion (USD 12 million); the sum of those two amounts.

Just to reiterate again, this is a ballpark figure. It might be twice, it might be half, but it’s probably not 10 times, or 10% of that amount.

And who are the beneficiaries of this transfer?

The beneficiaries of this cut will obviously be the people that use the toll roads. So, who are they, and how much benefit will they actually get each?

Poor people will get some of the transfer, but not much. Poor people either ride motorbikes or the train, so avoid toll roads, or ride buses or “travels”, where their share of the toll is between one fifth and one fiftieth of the fee, depending on the number of people that are in the vehicle. If you have packed full a Toyota Avanza driving from Jakarta to Surabaya, taking every toll road on the way, both ways, your toll fees are about IDR 500,000. So, assuming 5 passengers with bags, the cost per passenger is about IDR 100,000. So, the saving to each poor person is about IDR 30,000 (USD 2.25). If the trip is shorter, or the vehicle is bigger, the savings will be smaller.

Even to a poor person, this is not much. A bus ticket to Surabaya around this time of year is around IDR 500,000, so you might be saving them 6% of their fare. This probably doesn’t even cover their food costs for the journey.

And, this is assuming that bus and travel companies adjust their prices at all. There is a lot of research that says that prices are sticky, so most likely the transportation companies are the ones that will get the windfall profit.

So, other than some poor people and transportation companies that serve them, you’ll have middle-class and rich people, who care even less about a savings of IDR 100,000 on their carload than poor people do about their IDR 30,000.

So, with this policy, the government extracted around USD 12 million from toll road operators, and spread it around a bunch of people that didn't really need it and won't even really notice it.

So why is this such a bad thing?

So, why am I making such a big fuss about this? How does costing toll road operators USD 12 million, most of which is borne by an SOE anyway, cost the government so much money? I’m going to deal with three reasons, as follows:

1. It’s not fair;
2. It’s operationally disruptive; and
3. It sets (or reinforces) terrible regulatory expectations.

It’s not fair

There will be a certain group of people that will scoff at this: “Billion dollar companies can’t afford to give the average citizens a measly tariff cut for a few weeks?”

It’s true, the amounts are relatively small for a billion dollar company, but the point isn’t the amount, the point is that the government made a deal, and signed a contract, now they are unilaterally renegotiating it.

If we imagine a toll road which, in present value terms, will cost USD 1 billion, where the revenues will cover USD 900 million of that. To get it done, the government will have to pay a USD 100 million subsidy and the cashflows will look like the below, in present value terms.

When bidders bid on infrastructure projects in open tenders, they are bidding pretty close to the subsidy (tariff, concession fee, whatever the bid factor/balancing item is) that will exactly cover their costs and no more. They have to work as hard as they can to push down their capital and operational costs, put together the best financing package, convince their lenders to lend at the lowest rate possible, design the best/cheapest systems, and generally do everything they can to bid as cheaply as possible while meeting the minimum service standards. To win, you’re really cutting close to the bone.

So, any tariff cut or unexpected increase in costs, even a small one, means the investors will have lost money.

Imagine you are having lunch with some friends, then the President turns up and says “Surprise! Everyone’s lunch today is free!” then walks out, leaving the restaurants to bear the cost of your free lunch. This is pretty much what is happening here.

The concessionaire agreed to provide a service and bid their tariff according to a particular revenue projection, and all of a sudden, they’re being told that they aren’t allowed to charge what it says they are allowed to in the contract. That’s just not fair.

It’s operationally disruptive

The government has instructed an operator to cut tariffs; how much cash have they got in the bank?

If they have optimised their system, they’ll have as much as they need for that month and not a rupiah more. If they have any excess cash, they should use it to pay off their debts or dividends to their shareholders. All of a sudden, they’re being told that they’re going to lose 25% of their revenues for two weeks with no notice, and the banks are all about to go on holiday.

If a particular toll road operator loses USD 1 million, that means they need to scramble around to find this extra USD 1 million so that they can pay their employees’ salaries and keep the lights on. It’s not a massive amount of money for them, but it’s not easy to mobilise at short notice.

Toll road operators will also need to figure out how to implement a tariff cut with no notice. Tariff adjustments are things that they only expect every 2 years according to a schedule, so they have time to print up new signs, update their computer systems, figure out how much of every denomination of cash they stick in each toll booth, and so on. They, most likely, aren’t set up to change tariffs at a moment’s notice for a 16 day period.

Finding cash and implementing a no-notice tariff cut are not insignificant operational disruptions. Dealing with them will take significant time from the senior management, and probably the lenders as well. The cost of these disruptions is not included in the USD 12 million figure I estimated above.

It sets (or reinforces) terrible regulatory precedent

By cutting tariffs like this for one year in Lebaran, Indonesia has now created an expectation on the part of bidders that they will do the same in future. In Indonesia, for the next couple of years, faced with the same projections of cost and revenues, bidders will bid, discounting their revenues by 1.1%, accounting for the expectation that the government will force them to cut their revenues. For our billion dollar toll road above, going forward in Indonesia, the cashflows in present value terms will now look like the below.

The government will have to pay another USD 10 million in subsidy to cover the cost of the expectation that they created by cutting tariffs like this.

So how much will this cost the government in the long-run?

The government has 5 toll road projects in the “prospective” category of BAPPENAS’s priority project list, with capital costs as follows:

Manado – Bitung – USD 353 million

Tanjung Priok Access – USD 612.50 million

Balikpapan – Samarinda – USD 1.2 billion

Cileunyi – Sumedang – Dawuan – USD 1015.8 million

Pandaan – Malang – USD 420 million

The total capital cost of these toll roads is USD 3.6 billion. Assuming a capex/opex split of 60/40 and that land acquisition cost will be borne by the government, the total net present cost of these toll roads to the private sector will be on the order of USD 6 billion. Assuming no VGF, 1.1% of revenues needed to cover these toll roads represents USD 66 million.

So, future bidders in the toll road sector will “price in” the expectation that the government will force them to cut their tariffs during Lebaran. This price will be passed through to government through requesting a higher subsidy, or paying a lower concession fee; or through to users by charging a higher tariff.

In addition to the USD 66 million, investors will look at a government that is willing to surprise them with multi-million dollar losses not envisioned in their contracts, and expect that there will be a few more surprises as well. To cover this expectation, they will need to build an even larger margin of safety. It is eminently plausible that the margin the investors will assign to this uncertainty will be on the order of USD 100 million dollars.

The broader impacts

The expectation of regulatory intervention applies outside the toll-road sector as well. A government that is happy to force toll road companies to cut their tariffs might force private water operators to freeze their tariffs, apply arcane leverage constraints to airlines, or dramatically change the venues through which they can sell their product.

A government that acts erratically in its regulation is a significant driver of the country risk premium that investors and lenders will apply when deciding how to value their products in Indonesia. This flows through to the cost of finance, which is a component of almost every good in the country pushing up costs to all consumers, and harming the competitiveness of Indonesia's firms, relative to those in countries where the government does not take actions like this.

So, what can they do?

The damage isn’t necessarily completely done. If the government compensates the toll road operators for the gap between the tariff they normally charge and the one they’re being forced to charge, that would mitigate a lot of the impact of the expectations of regulatory meddling.

Even if they don’t do that, they can still mitigate some of the poor expectations by the way they design their concession agreements going forward. All they need to do is build in the right to cut tariffs whenever they want, but make it clear that they must give reasonable notice, and that the government will meet the gap between their agreed tariff, and the tariff that the government may choose to impose.

It will be interesting to see how the next round of tenders proceeds following this announcement.

New course on PPPs in developing countries

The World Bank is about to launch a new course to help explain how public private partnerships (PPPs) can help bridge the infrastructure gap and fight poverty.

You can register for the course for free here through Coursera. It runs for 4 weeks, starting from 1 June.