STUMP » Articles » Public DB Pension Fund Managers, Why Don't You Split Out Cash Flows? » 21 November 2017, 04:04

In yesterday’s post, I highlighted a central graph which supposedly shows that public pensions are trotting along just fine, thank you very much.

This is the graph:

The central argument is that in almost all the fiscal years, the cash flows from contributions and investment income are enough to cover benefits.

As I noted, investment income is supposed to be covering benefits (not necessarily solely – there is nothing wrong with having to liquidate assets to cover retiree benefits. That is, if the obligations aren’t growing. Let’s come back to that in a later post.)

But let us humor the people of NCPERS who wrote this report.

Before I get into mine, John Bury’s take on the NCPERS paper. I absolutely agree with Bury’s title:

NCPERS Looks To Dismantle Public Pensions

I guess NCPERS needs to remember why pensions need to be pre-funded.

MATCHING THE ORIGINAL GRAPH

The data came from the U.S. Census Bureau, which combines some state and local pension info. As noted by NCPERS, it covers 1993 – 2016 (fiscal years). I have saved my spreadsheet with the underlying data here.

First, let me reproduce their line graph from the data I grabbed, so you can see we’re working from the same data. Here is my reproduction:

I noticed that our data differ for 2015, and that may be a timing issue. I still see some weird stuff in the 2015 data sets, and there may be additional data clean-up that NCPERS did to fix double-counting. That said, the differences between the NCPERS graph and mine are minimal, and I don’t think they screwed around with the data.

They just screwed around with what they displayed. I will explain as I give alternative graphs from the same data.

ALTERNATIVE GRAPHS

I’m going to do multiple alternative graphs. First, I will be reproducing the line of benefits paid (no problem), and doing a stacked graph of the contributions and the investment income. Sound good?

Well, some years, investment income was negative. That causes a problem for doing stacked graphs, so I will diminish the contributions by that amount. When investment losses exceed contributions (2002, 2009), I will have all the negative be attributed to the investment income. Don’t worry, this will become even simpler soon.

Here is alternative graph 1:

A few things to note:
- When I remove assets from the graph, the scale comes down, so you can better see how much contributions + investment income deviates from benefits paid that year. The main reason to put a balance sheet item on the same graph as a cash flow item is to make the cash flow deviations look small.
- Whoa, that investment income sure does fly up and down a lot, which is a bit odd for a very steadily increasing line of benefit payments.
- The fluctuations of investment income tend to swamp the contribution amounts, either way.

That was hard to look at, and as I said above, contributions are supposed to cover future benefits, not current ones. So let’s compare investment income against benefits.

Alternative graph 2:

Ew, that looks really bad. That said, this is where NCPERS’ “cushion” argument comes into play. Let me address that cushion.

WHAT HAPPENS IN BAD INVESTMENT YEARS

I am going to excerpt all the references to “cushion” in the NCPERS paper below.

But this is the gist: in bad investment years, you will have assets that built up in good years that you can sell to cover benefit flows.

This is also known as buy high, sell low. When investment income is negative, one needs to be selling assets… in a down market. The reason investment income is so volatile is that so much of public pension funds are in volatile assets. (That’s for a later post).

With regards to a very steadily growing benefits line – that’s cash flows, not accrued liability – and extremely volatile investments, that is plainly not sound risk management.

Here are all the cushion quotes from the the NCPERS paper.

[page 2]
….pension funds ability to pay benefits depends on two things: (1) contributions and investment income is more than benefit obligations in a given year and (2) there is a sufficient cushion to weather an economic recession.

[page 3]
In cases where expenditures exceeded income, we examine whether the pension funds had enough cushion to meet the shortfall. (By “cushion” we mean the sum of money that is accumulated when income exceeds expenses.)
….
But in the remaining 20 years, when income from contributions and investment earnings was more than benefit obligations, pension funds were building up a cushion that enabled them to weather the 2001 recession, the Great Recession of 2008, and other economic downturns.

[page 4]
They had built up enough of a cushion during normal times that they were all able to meet their pension obligations. Today, state and local pension funds sit on a huge pile of money, about $3.9 trillion, that will provide a cushion during future economic recessions.

[page 5] Our analysis shows that despite rising liabilities during the last quarter century, pension plans have been able to meet their annual benefit payments from contributions and investment income due to the cushion they built up in good years.

[page 8]
Column 4 shows the cushion built up during all the years in which contributions and investment income exceeded benefit payments, and column 5 shows the plans’ total assets.
…..
States in both top- and bottom-funded groups on average experienced situations in which contributions and investment income was not enough to meet annual benefit obligations about six out of 24 years during 1993 2016. The cash flow shortfalls were caused by the 2001 and 2008 recessions as well as other economic downturns. But both types of funds – partially and almost fully funded public pension plans – had adequate cushions to cover the cash flow shortfall.

[page 10]
This is because there are usually more good years than bad years as far as investment returns are concerned, and good years allow pension funds to build up a cushion that can cover the shortfall during recession years.

Policy Options

State and local pension funds are resilient. They have weathered some serious economic storms in the last quarter century. Their assets now exceed pre–Great Recession levels. They have been able to meet their benefit obligations even in recessions due to the cushion they built up during normal and good economic times.

[page 11]

As long as pension plans’ contributions and investment income exceed their annual benefit obligations, the plans can be sustained in perpetuity. This is because more often than not, pension funds’ income surpasses their expenses, which allows them to build up a cushion to pay benefits during economic downturns.

Notice that last bolded clause?

ANYBODY NOTICE WHAT A HUGE ASSUMPTION THAT IS?

via GIPHY

One may like big assumptions, but this one can have devastating effects.

STOP ASSUMING THINGS THAT MAKE PENSIONS MORE LIKELY TO FAIL

Here are two specific posts that relate to this:
Post 1: The Fragility of Can’t Fail Thinking

Because they thought that pensions could not fail in reality, that gave them incentives to do all sorts of things that actually made the pensions more likely to fail. Because, after all, the taxpayer could always be soaked to make up any losses from insane behavior.

Here is what I see as insane behavior — pension obligation bonds (adding leverage) and chasing yields (making investment income even more volatile).

Post 2: Public Pensions: Why Do 100% Required Contribution Payers Have Decreasing Fundedness?

SOMEONE WILL PAY….LATER….MAYBE

Many plans are using approaches to determining the ARC – in both the normal cost and amortization of the unfunded liability – that will drive the funded ratio to decrease even when economic conditions are good.

It’s not even a matter of the most egregious practices, like spiking, early retirement inducements, DROP benefits, and retroactive granting of benefits. It’s that inherently the valuation approaches always assume somebody will pay more tomorrow to make up for losses.

That is the assumption. That there will always be contributions made by people later to cover for shortfalls.

And if there isn’t?

Talking about perpetuity when it comes to human systems seems not very realistic to me.

FINAL ALTERNATIVE GRAPH

So one last graph, because we’ve seen: investment income is volatile, benefits are steadily growing (about 5% per year), so… what’s happening with contributions?

Hmmm, contributions are less than benefits. Again, not necessarily concerning — IF sufficient assets were contributed back when retirees were working, producing reliable investment income now.

Um, right, reliable investment income.

I will look more closely at these contributions in the next post.

Spreadsheet with data and graphs.