Pension Shocks and Wages

Transcription

1 Pension Shocks and Wages Pawel Adrjan* & Brian Bell** *UniversityofOxford ** King s College, London and CEP October /33

2 Structure of Talk Pension Background Funding and Deficit Payments Theoretical Considerations Data Results Conclusions & Future Work 2/33

3 Key Points 15-20% of compensation comprises non-wage elements. How do firms respond to shocks to these elements? Our context - a large rise in deficit payments by UK firms to cover pension black holes Many firms have also closed their schemes to workers. We link to individual panel data on wages for workers in these firms. Workers bear some of the burden of the deficit costs - if they are (or were) members of the scheme. Firms make significant immediate cost savings when they close these schemes - but exits rise. 3/33

4 Pension Background Most large private firms in the UK used to o er Defined Benefit (DB) pensions to workers. During the 2000s, these schemes increasingly closed down and were replaced with Defined Contribution (DC) schemes. By 2016, 85% of DB schemes were closed to new members, and 35% were also closed to future accrual. 4/33

5 Decline in DB schemes Figure: Active Member of DB Schemes, /33

6 Pension Background There are currently 5,886 DB schemes with assets and liabilities in the UK - with 10.9 million members. As a result of the closure of schemes, active members make up only 13% of the membership of these schemes. The schemes have total assets of 1.4trn, and liabilities of 1.6trn (or 2.1trn on a full-buyout basis). 6/33

7 Funding It is the legal responsibility of the firm to maintain the funding of the DB scheme - until the last member dies or the scheme liabilities are transferred. The firm can only avoid this by declaring bankruptcy - but unlike the US, they cannot use Chapter 11-type arrangements. Crucially, the firm does not decide on the funding level - that is up to the independent trustees of the scheme. Firms can cut their losses by transferring risk to an insurer but this is expensive - liabilities are valued at a much lower discount rate upon a buy-out than for funding purposes. 7/33

8 Funding Every three years, the trustees must obtain a formal valuation of the assets and liabilities of the scheme from an actuary. If the scheme is in deficit, the trustees must produce a recovery plan to restore the funding level - this is subject to approval by the Pension Regulator. Since most of the accrued benefits are irrevocable, the recovery plans focus on increased contributions from the firm ( deficit payments ). These are mandatory payments over and above normal employer contributions that relate to current and new members - which the firm can control to some extent. 8/33

9 Funding If the firm becomes insolvent and the scheme remains in deficit, the DB scheme enters the Pension Protection Fund (PPF). The PPF is funded by annual levies on all DB schemes. Generally under PPF the compensation is (i) full for retired members, (ii) full for ill-health retirees, (iii) 90% for deferred and active members, subject to 37k cap. Deferred and active members therefore gain from deficit payments to the extent that it keeps the scheme funded and entitles them to 100% of benefits at retirement. Risk is that it sends the firm into insolvency. 9/33

10 Normal vs. Deficit Contributions Normal0 Contributions Special0Contributions Figure: Normal and Deficit Payments into DB schemes, ONS 10 / 33

11 Rising Share of Labour Costs Going to Pensions Figure: Components of total compensation of employees, (UK national accounts) 11 / 33

12 Closing the Scheme Firms can close the DB scheme to (a) new members (soft freeze) or(b)toexistingmembers(hard freeze). May o er DC as an alternative - and must do so since auto-enrollment. Protects firms from future additional deficit payments - and may bring immediate savings. No restriction on firms doing this in theory. 12 / 33

13 Theory In standard frictionless competitive model, deficit payments have no e ect on wages in the firm. Firm is a price-taker for workers and faces a perfectly elastic supply curve. If the deficit payments yield a benefit to the worker, then there can be a wage fall - firms are actually price-takers in a total compensation sense. Closing schemes should also generate no immediate saving - lower employer contributions should be o set by higher wages. 13 / 33

14 Theory Rent-sharing model (or a model with frictions) can predict a hit to wages from the deficit payments. Firm profits are: p = R(p) wl rk C Workers and firm engage in a Nash Bargain with standard maximization problem: max [qln {[u(w) u(w)] n} +(1 q)p] 14 / 33

15 Theory First-order condition implies that: w = w + q 1 q p n Since C directly a ects profits, it reduces wages all else equal. Size of the e ect depends on relative bargaining strength. 15 / 33

16 Related Literature Literature on the impact of pension contributions on investment (Rauh (2006), Liu and Tonks (2013)) Very little on wages. Rauh et al (2017) show that US firms that freeze their DB plans subsequently reduce their payroll costs substantially. Similar in many ways to the Gruber (1997) Chile payroll tax paper - legal incidence falls on the employer, but economic incidence need not. 16 / 33

17 The Data Our universe of firms are the largest 300 (by market cap) on the London Stock Exchange between , domiciled and registered in the UK. There are 475 firms - accounting for about 95% of UK market capitalization. Roughly equivalent to FTSE350. For each firm, we hand-collect annual data from the Report on UK DB pension exposure - required by IAS19. We have annual contributions, current service cost, assets and liabilities, scheme closure date. 17 / 33

18 The Data Of the 475 firms, 65% have a DB scheme (either open or closed). 29% of these close their scheme during our sample period. We link the firm to workers in ASHE using the firm identifier. This can be done only for the period however IAS19 (amended) only came into e ect in 2001 so this is not a major constraint. We have about 50k workers who work for a firm with a DB scheme. 18 / 33

19 Data Comparability Total&DB&Employer&Contributions ONS.DB.Normal.+.All.Special.Contributions Sample.DB.Total.Contributions.(rightFhand.axis) Figure: Our Sample v Total ONS Contributions 19 / 33

20 Deficit Payments Firms do not have to report deficit payments separately in the accounts (though some do). We use the following equation to estimate deficit payments: Deficit Payment = Total Employer Contributions - Current Service Cost IAS19 defines current service cost as the increase in the present value of a defined benefit obligation resulting from employee service in the current period. 20 / 33

21 Estimated v Actual Deficit Payments 0 million 2,000 4,000 6, Year DB deficit contributions (employer contributions - current service cost) DB deficit contributions (company reports) Figure: A comparison where firms do report deficit payments (n = 111 firms) 21 / 33

22 Empirical Framework We estimate a standard fixed-e ect panel for wages: w ijt = a i + b j + g t + Â d k Deficit jt k + qx ijt + # ijt for worker i at firm j at time t To make the deficit measure comparable across firms, we deflate by the initial level of employment. Standard errors are clustered at the firm-level. When looking at scheme closure, we include a dummy variable. 22 / 33

23 Deficit payments per worker Fraction Annual deficit payments ( 000 per employee) Figure: Deficit payments divided by initial employment level 23 / 33

24 Results - All Workers (1) (2) (3) db_emp(-1) (0.049) db_emp(-2) * (0.099) * (0.052) * (0.105) (0.055) ** (0.105) db_emp * (0.139) * (0.149) * (0.154) Individual FE x x Year FE x x x 5-digit Industry FE x Firm Performance Controls x x x Firm FE x Match FE x 1-digit Industry * Year FE x x x Sample Size 171, , , / 33

25 DB exposure We can define three mutually exclusive groups of workers within these firms: Never a member of the DB scheme (with some small error) A deferred member - the scheme has closed to future accrual (or the worker has chosen to opt out) A current member of the DB scheme 25 / 33

26 Results - DB membership type (1) (2) (3) db_emp * DB never member (0.092) (0.101) (0.091) db_emp * DB deferred member (0.150) * (0.160) * (0.171) db_emp * DB current member ** (0.205) ** (0.021) ** (0.224) Estimated Sharing Rate 9.2% 10.0% 10.3% Individual FE x x Year FE x x x 5-digit Industry FE x Firm Performance Controls x x x Firm FE x Match FE x 1-digit Industry * Year FE x x x Sample Size 171, , , / 33

27 DB Closure We have 91 firms in our sample that close their DB scheme over the period We estimate the impact on employer contributions and wages of these closures. Initially assume the closure is exogenous. Then use an IV approach based on accounting deficit of pension scheme - plausibly unrelated to current economic environment of firm. 27 / 33

28 Results - DB Closure on Employer Contributions (1) (2) (3) (4) Scheme Closed (1.877) (1.675) (1.160) (0.731) Scheme Closed * DB member *** *** * (2.372) (1.729) (1.459) db_emp (0.209) (0.175) (0.152) (0.152) db_emp * DB member (0.243) (0.169) (0.170) First-Stage F Stat 17.7 Match FE x x x x Year FE x x x x 5-digit Industry FE x x x x Firm Peformance Controls x x x x 1-digit Industry * Year FE x x x x Sample Size 95,022 95, , , / 33

29 Results - DB Closure on Wages (1) (3) (4) (6) Scheme Closed * (0.012) (0.012) (0.012) (0.010) Scheme Closed * DB member * (0.021) (0.021) (0.014) db_emp * (0.151) (0.091) (0.091) (0.095) db_emp * DB member ** ** * (0.197) (0.197) (0.199) First Stage F-Stat 17.7 Match FE x x x x Year FE x x x x 5-digit Industry FE x x x x Firm Peformance Controls x x x x 1-digit Industry * Year FE x x x x Sample Size 171, , , , / 33

30 Worker Mobility The results suggest significant frictions that allow DB firms to make savings from scheme closure. An obvious explanation is that DB workers are less mobile than other workers - likely to be increasingly true in an environment in which alternative employers do not o er DB schemes. Potential cost for the firm in closing the scheme is that this unshackles the worker and increases exit - particularly problematic if the exits are from the top of the ability distribution in the firm. 30 / 33

31 Results - Worker Mobility (1) (2) Panel A: All Workers in ASHE Ever a DB Member *** (0.000) *** (0.000) Sample Size 1,033,297 1,018,291 Panel B: Firm Sample Ever a DB Member *** (0.003) *** (0.002) Sample Size 216, ,482 Panel C: Scheme Closure Scheme Closed if Never a DB Member 0.068*** (0.008) Scheme Closed if Ever a DB Member 0.125*** (0.019) Scheme Closed if Current DB Member 0.157*** (0.019) 0.090*** (0.008) 0.148*** (0.019) 0.199*** (0.021) Sample Size 216, , / 33

32 Conclusions Wages fall in firms that are required to make DB deficit payments - but only for those workers exposed to the scheme. Firms make substantial immediate savings when they close a DB scheme to future accrual. Latter result is inconsistent with a frictionless labour market. Evidence suggests that scheme closure significantly increases exits from the firm. 32 / 33

33 Future Work When scheme closes, mobility increases for those who were in scheme - may reduce gains for firms? Small samples, but preliminary work suggests that exits are more prevelant amongst those toward the top of the earnings distribution. Alternative instrument - date of formal actuarial valuation - surely exogenous since determined historically. 33 / 33