UK – Detailed research from JPMorgan Asset Management (JPMAM) into the effect of quantitative easing (QE) on pension fund solvency suggests the first bout of bond buying might have reduced the average UK pension scheme’s funding level by 16% – but that QE2 had almost no effect at all.

While QE appears to have lowered 10-year Gilt yields from where they would otherwise have been on both occasions, the effect at the longer end of the yield curve became much more muted during QE2.

By modelling the entire liability curve of an average UK pension scheme, JPMAM show that, because its liabilities are so long-dated, the impact on overall solvency was greatly diminished the second time around.

QE has become a contentious issue for the UK pensions industry.

The National Association of Pension Funds (NAPF) called for “more of a debate” on its impact on pension schemes in July last year, and most recently voiced its concerns about “unintended consequences” in a January 2013 submission to the Treasury Select Committee’s enquiry on QE.

The JPMAM research begins by building a model for counterfactual nominal and real yield curves, based on 12 variables including the Bank of England base rate, the Federal Reserve funds rate, the VIX index, US non-farm payrolls, the UK index of production, UK unemployment and other macroeconomic factors.

Plotting the counterfactual 10-year yields against what actually happened in bond markets, the results show that the Bank of England’s Inflation Report early in 2009, which “telegraphed” QE, led to a marked divergence as the counterfactual rate spiked and the actual rate fell.

Between the end of QE1 and the implementation of QE2, the two rates converged again, before diverging significantly as the Bank of England entered the market for the second time.

Paul Sweeting, European head of JPMAM’s strategy group, said: “QE appears to have had a significant depressive effect on interest rates, and QE does appear to have been the main culprit for this.

“But while there has already been work on the impact of QE on rates, there has not been anything detailed looking at the impact on pension schemes, especially looking at the full period over which QE has been in place.”

To build a model of the average pension scheme’s assets and liabilities, JPMAM took the Pension Protection Fund’s (PPF) monthly reporting of solvency levels for covered schemes, and extrapolated the average scheme’s solvency, with an estimate that funding solvency is, on average, 94.2% of PPF solvency.

The firm then built a population based on UK pension fund mortality and a mix of fixed and real liabilities that correlated as closely as possible to changes in PPF liabilities.

With a complete cash-flow curve for the average scheme’s liabilities, JPMAM was able to estimate the effect on the discount rates for the points on that curve using the counterfactual and actual curves for nominal and real bond yields.

The results suggest QE1 reduced the average scheme’s solvency by 16% – the equivalent of a £150bn (€176bn) hike in aggregate UK pension deficits.

JPMAM list a number of caveats for this result – including the fact it did not take account of the impact of QE on risk assets, due to the complexity of the relationships involved.

However, Sweeting pointed out that the widely cited estimate that QE1 boosted equity valuations by 20% would still not have cleared the estimated excess deficit due to QE.

More important, the research finds that, while the divergence between the actual and counterfactual yields was high during QE1 and became even higher – more than 200 basis points higher – during QE2 at the five and 10-year points on the nominal curve, the divergence fell at the 20-year point on the curve and had become almost negligible by December 2012.

A similar result holds on the real yield curve.

Because pension fund liabilities tend to be long-dated, that meant that, while the model scheme’s funding solvency significantly worsened relative to the counterfactual solvency level during QE1, it was actually higher than the counterfactual solvency level between QE1 and QE2 – and QE2 itself appears to have had very little effect at all.

By December 2012, the actual solvency level was almost perfectly in-line with the counterfactual solvency; in other words, the impact of QE has apparently completely worn off.

“The potential impact of QE [on pension scheme funding] was over-emphasised during QE1,” said Sweeting.

“It was a fairly short period over which QE had an impact. The overall impact on pension schemes is much lighter than it might at first appear if you only look at the 10-year yield.”

Sweeting told IPE he and his team would be taking the results of their research for discussion at the Bank of England next week.