The purpose of this article is to investigate the technical efficiency of Indian pension funds using data envelopment analysis (DEA) and to assess the reasons of inefficiency if any. This article analyzes the efficiency performance of all the pension funds available to Indian subscribers from the year 2015–2019 using radial measurers (BCC) of DEA based on secondary data collected from the annual reports of New Pension System Trust. Findings indicate that the average efficiency of Indian pension funds was 75.38% and this sector experienced overall stability in the average efficiency levels during the study period. Almost 40% of the pension funds operated efficiently in one or more years during the study period. The minimum slack was found in the input “expense ratio.” This confirms that risk associated with investments is the cause of inefficiency and not the expense ratio in the Indian pension sector. Tobit regression is applied to explore the main drivers of efficiency in the India pension funds. The study finds that fund size has positive association with the efficiency of the pension funds. Public sector funds were more efficient than the private sector funds. This study is first of its kind that has assessed the efficiency of Indian pension funds. The study brings into light the operating characteristics and efficiencies of the Indian pension funds for the period 2015–2019 and therefore holds important insights for policy makers, practitioners, and decision-makers.

1 INTRODUCTION

The pension sector in India has great potential and is one of the fastest-growing sectors of the Indian financial system. India has been a late starter in introducing a regulatory framework and a universal pension plan. Indian pension sector has mainly covered organized sector, almost 10% of the Indian population that means there is tremendous potential in the Indian pension sector. Initially, the Indian pension industry offered “defined benefit” (DB) pension plans to the central government (CG), state government (SG), and public sector employees. DB pension system was stressing the Indian fiscal system and there was also need to provide pension schemes in the unorganized sector of the economy. Therefore, the government of India established pension funds regulatory and development authority (PFRDA) in the Indian pension sector. PFRDA introduced a new pension scheme (NPS) in 2004 for government employees in India. Later on, it was opened for all the sections of the society in the year 2009. NPS is now open for both organized as well as unorganized sector.

The NPS achieved impressive growth of 35.65% in their asset under management (AUM) during the year 2018–2019 as compared to the previous year 2017–2018. All the pension funds recorded double-digit growth. The detail of AUMs under various schemes is given below in Table 1. As is evident from Table 1, AUM of CG, SG, Atal Pension Yojana (APY), and Scheme A Tier I (A1) schemes posted high growth of 28.32%, 36.98%, 79.69%, and 198.77%, respectively.

TABLE 1. Asset under management (AUM) under various schemes (amount in Rs million; 1 million = Rs. 1,000,000)

Sr. no.	Scheme	2015	2016	2017	2018	2019
1	Equity Tier I (E1)	6545.0	11,814.6	25,389.8	43,082.2	72,342.1
2	Equity Tier II (E2)	435.4	604.4	1259.0	2177.8	3254.4
3	Bonds Tier I (C1)	4687.6	8877.8	16,849.5	28,465.5	44,220.7
4	Bonds Tier II (C2)	374.8	551.0	1013.3	1621.6	2090.8
5	Govt. Sec Tier I (G1)	7712.9	13,247.9	25,069.3	42,430.6	68,967.5
6	Govt. Sec Tier II (G2)	356.8	543.5	1124.3	1814.7	2626.9
7	Scheme A Tier I (A1)	—	—	10.4	65.3	195.2
8	Scheme A Tier II (A2)	—	—	1.0	—	—
9	NPS Lite	16,057.2	21,075.5	26,392.1	30,058.2	34,092.3
10	Corporate CG (CCG)	41,051.2	68,050.5	10,7534.7	14,8463.3	20,6828.3
11	Central Govt (CG)	36,7367.7	48,1347.8	67,0400.5	84,9546.0	1,09,0107.0
12	State Govt (SG)	36,3962.6	57,6925.0	85,1714.3	1,15,9884.8	1,58,8811.1
13	Atal Pension Yojana (APY)	—	5063.4	18,850.0	38,178.6	68,603.0
Grand total		808,551.2	1,188,101.4	1,745,608.2	2,345,788.6	3,182,139.3

Source: Annual Reports NPS Trust.

This huge corpus of fund is managed by the following fund managers. The list of fund managers and AUM is given below in Table 2. All fund managers continued to register impressive growth in AUM. The pension sector in 2019 recorded growth of 35.65% in the AUM as compared to 2018. Being the new entrant in the pension market, Birla Sunlife Pension Management Ltd. recorded the highest growth of 285% in AUM in the year 2019 (NPS Trust annual report, 2018–2019).

TABLE 2. Asset under management (AUM) with the pension fund managers (PFM) (amount in Rs million; 1 million = 1,000,000)

Sr. no.	PFM	2015	2016	2017	2018	2019
1	SBI Pension Fund Pvt. Ltd.	314,071	460,188	667,232	8,92,832	12,19,590
2	UTI Retirement Solution Ltd.	248,314	359,182	520,431	7,01,302	9,27,193
3	LIC Pension Fund Ltd.	240,101	355,119	527,093	6,94,832	9,37,077
4	ICICI Prudential Pension Funds Management Company Ltd.	3690	7011	14,415	25,603	51,647
5	Kotak Mahindra Pension Fund Ltd.	1074.744	1727	3120	23,255	34,760
6	Reliance Capital Pension Fund Ltd.	769.6571	1112	1690	5362	7847
7	HDFC	530.7745	3762	11,630	2310	2893
8	Birla Sunlife Pension Management Ltd.	—	—	—	294	1132
Total		808,552	1,188,102	1,745,610	23,457,90	31,821,39

Source: Annual Reports NPS Trust.

2 REVIEW OF LITERATURE

Very few studies have examined the efficiency of the pension funds using data envelopment analysis (DEA). First such study was conducted by Braberman et al. (1999) on Argentinean pension funds. They used Cobb–Douglas frontier and concluded that although regulations increased cost but did not affect the efficiency of the funds significantly. Radial models variable return to scale (VRS) and constant return to scale (CRS) were used by Barrientos and Boussofiane (2005) to assess the efficiency of Chilean pension funds during the period 1982–1999. They found variation in the average efficiency levels during the period. They also found positive relationship between market shares and efficiency scores. To assess the effect of mergers and acquisitions on the efficiency of Portuguese pension fund companies during the period 1994–2003, Barros and Garcia (2006) used BCC, constant return to scale (CRS), cross efficiency, and super efficiency measures and found high level of average technical and scale efficiencies. They also found that merger and acquisitions have positive significant relationship with the efficiency of the firms involved in it. Bikker and Dreu (2006) in their study on Netherlands' pension funds found that industry-wide pension funds were significantly more efficient than company funds and other funds. Barros et al. (2008) studied the efficiency of 10 pension funds in Argentina using stochastic frontier model and found that the mean technical efficiency was 0.86.

To assess the efficiency of Australian pension funds during the period 2005–2009, Sathye (2011) used radial production model along with the Tobit regression. He found low efficiencies in the pension funds during the study period. Using the Tobit model, he found that fund size and investment in risk-free assets positively affected the efficiency whereas “number of products offered” negatively affected the efficiency. Andreu et al. (2014) evaluated the efficiency of the Spanish pension funds using slack-based measures of DEA. They evaluated funds categorically and found Variant III the most efficient. Galagedera and Watson (2015) evaluated Australian pension funds of four categories, namely, corporate, retail, public, and industry. They found that membership and proportion to risky assets had a negative association with the fund performance but fund size had a positive association. Drazenovic et al. (2019) evaluated Croatian pension funds for the period 2015–2018 using CCR and BCC models of DEA. They found that Croatian pension funds were fairly efficient. Lin et al. (2020) evaluated investment trust corporations in Taiwan using the additive network DEA approach for the period from 2011 to 2017. They found that overall mean technical efficiency was 0.555 with an 18.3% variation in efficiency scores. The present study contributes to the literature on pension funds efficiency in several ways. First, we assess the efficiency of the Indian pension funds. Second, we explore the reasons of inefficiency and third, we explore the main drivers of efficiency.

3 OBJECTIVES

There are three main objectives of this study. The first objective is to assess the efficiency of Indian pension funds. The second objective is to find the reasons of inefficiency. Finally, the third objective is to explore the main drivers of efficiency and check whether it confirms or contrast the past findings.

4 DATA AND METHODOLOGY

Data for this study has been taken from the annual reports of the NPS trust, India for the period starting from the financial year 2014–2015 to 2018–2019. Here, financial year means a period from April 1 of the year to March 31 of the next year. Presently, there are total eight pension fund managers in the Indian pension sector, which are managing a total of 64 pension funds under 13 different fund categories. Majority of the studies on mutual fund efficiency evaluation have used mean returns as the output variable and risk (total or systematic), expenses, and minimum initial investment as input variables (Basso & Funari, 2001; Choi & Murthi, 2001; Morey & Morey, 1999; Sedzro & Sardano, 1999; Sengupta & Zohar, 2001; Sharma & Sharma, 2018). Return of funds is a conventional output in the DEA studies whereas conventional inputs are risk (standard deviation, beta) and the expense ratio (management fees, administrative expenses) (Daraio & Simar, 2006). Fund returns are annualized returns net of expenses. The standard deviation of returns shows the fund's total risk whereas the beta coefficient (systematic risk) represents the fund's volatility in compare to market return. Systematic risk cannot be reduced or removed even after diversification (Sharpe, 1966). Consistent with the previous studies, this study has used return as an output variable and standard deviation, beta, and expense ratio as the input variables. Our input and output variables for this study are as given below in Table 3.

TABLE 3. Descriptive statistics for pension funds (2015–2019)

Financial year		SD	Beta	Expense ratio	Return
2014–2015	Min	0.0353	0.3823	0.0134	0.0839
	Max	0.1561	1.2322	0.2632	0.1380
	Average	0.0811	0.6953	0.2037	0.1063
	SD	0.0402	0.2580	0.1077	0.0127
2015–2016	Min	0.0040	−0.1415	0.0134	0.0412
	Max	0.2553	0.9418	0.2632	0.1430
	Average	0.0882	0.2363	0.2161	0.0976
	SD	0.0580	0.4042	0.0987	0.0163
2016–2017	Min	0.0182	0.4174	0.0134	0.0855
	Max	0.1941	1.4013	0.2632	0.1660
	Average	0.0786	0.7050	0.2161	0.1060
	SD	0.0494	0.3480	0.0987	0.0141
2017–2018	Min	0.0234	−1.1193	0.0134	0.0647
	Max	0.1593	1.5333	0.2632	0.1549
	Average	0.0726	−0.2020	0.2062	0.0995
	SD	0.0401	0.7500	0.1057	0.0127
2018–2019	Min	0.0033	−0.3100	0.0134	0.0605
	Max	0.1380	0.8140	0.2632	0.1510
	Average	0.0604	0.2506	0.2125	0.0963
	SD	0.0385	0.2535	0.1013	0.0137

DEA is a linear programming-based technique for measuring the relative performance of organizational units where the presence of multiple inputs and outputs makes comparisons difficult. The DEA models include the CCR model, The VRS model, stochastic DEA, and nonparametric stochastic frontier estimation. DEA is a multifactor productivity analysis model for measuring the relative efficiency of a homogenous set of decision-making units (DMUs). For every inefficient DMU, DEA identifies a set of corresponding efficient DMUs that can be utilized as benchmarks for the improvement of performance and productivity. DEA is developed based on two scales of assumptions, namely, CRS model and VRS model.

Charnes et al. (1978) proposed a model, which had an input orientation and assumed CRS. Subsequent papers have considered alternative sets of assumptions, such as Banker et al. (1984) who proposed a VRS model. The following discussion of DEA begins with a description of the input-orientated CRS model that was the first to be widely applied. This study attempts to measure the technical efficiency using the input-oriented BCC model so it calculates the VRS (BCC) scores for the years from 2015 to 2019.

4.1 CCR and BCC input-oriented models

When using the ratio form of DEA, it becomes necessary to apply the BCC-DEA model. Suppose, there are n number of pension funds and each pension fund produces s outputs by using m inputs. Let $urn:x-wiley:14723891:media:pa2806:pa2806-math-0001$ be the amount of the ith input used by the kth pension fund, $urn:x-wiley:14723891:media:pa2806:pa2806-math-0002$ be the amount of rth output produced by the kth pension fund, $urn:x-wiley:14723891:media:pa2806:pa2806-math-0003$ be the weight given to the ith input of the kth pension fund, $urn:x-wiley:14723891:media:pa2806:pa2806-math-0004$ be the weight given to the rth output of the kth pension fund. Then, the efficiency of the kth pension fund can be represented as,

$urn:x-wiley:14723891:media:pa2806:pa2806-math-0005$

where, $urn:x-wiley:14723891:media:pa2806:pa2806-math-0006$

The CCR fractional program may be written as follows,

$urn:x-wiley:14723891:media:pa2806:pa2806-math-0007$

$urn:x-wiley:14723891:media:pa2806:pa2806-math-0008$

$urn:x-wiley:14723891:media:pa2806:pa2806-math-0009$

by normalizing the numerator of the above model, we get a multiplier model which is illustrated as below;

$urn:x-wiley:14723891:media:pa2806:pa2806-math-0010$

$urn:x-wiley:14723891:media:pa2806:pa2806-math-0011$

$urn:x-wiley:14723891:media:pa2806:pa2806-math-0012$

$urn:x-wiley:14723891:media:pa2806:pa2806-math-0013$

the dual of the above linear programming (envelopment form) which is also known as the input-oriented CCR model may be described as follows:

$urn:x-wiley:14723891:media:pa2806:pa2806-math-0014$

$urn:x-wiley:14723891:media:pa2806:pa2806-math-0015$

$urn:x-wiley:14723891:media:pa2806:pa2806-math-0016$

where, $urn:x-wiley:14723891:media:pa2806:pa2806-math-0017$ is unrestricted in sign, $urn:x-wiley:14723891:media:pa2806:pa2806-math-0018$ is non-Archimedean constant, $urn:x-wiley:14723891:media:pa2806:pa2806-math-0019$ is the dual variable corresponding to the jth constraint and is known as intensity variable, $urn:x-wiley:14723891:media:pa2806:pa2806-math-0020$ is the slack in the ith input of the kth pension fund, and $urn:x-wiley:14723891:media:pa2806:pa2806-math-0021$ is the slack in the rth output of the kth pension fund. On imposing the condition $urn:x-wiley:14723891:media:pa2806:pa2806-math-0022$ in the above model, it becomes the input-oriented BCC model.

Different frontier models have been used for the evaluation of DMUs. The parametric and nonparametric approach has been used for comparing performance by the researchers. Mogha et al. (2014, 2016) evaluated the technical efficiency of private and public sector hospitals of India using DEA-CCR and BCC output-oriented models. Mogha (2020) evaluated the performance of academic departments of selected private institutions in India using DEA- based dual CCR model.

The fundamental property of DEA is that it is independent of the units of the input and output variables, that is, unit invariant (Russell, 1988). According to Pastor and Lovell (1995), the BCC-DEA model is scale invariant in input or output variables but not in both. Hollingsworth and Smith (2003) found that the nature of available data necessitates the use of ratios, as ratios represent more accurately the production function rather than absolute numbers in DEA. They recommended the application of the BCC-DEA model when ratios are used as input and output variables. So, because of the nature of available data, this study has used the radial (BCC-DEA model, Banker et al., 1984) model of DEA. Efficiency estimates are calculated using deaR software, R version 3.6.

4.2 Specification of the regression model

The technical efficiency (VRS) scores of the pension funds estimated by DEA are used as dependent variable. Since efficiency scores lie between 0 and 1, therefore Tobit regression model is suitable for this study. The Tobit regression model has also been used in similar studies by Barrientos and Boussofiane (2005), Sathye (2011), and Galagedera and Watson (2015) Independent variables for this study includes AUM, age of the pension funds, sector (government or private) and type of fund manager (public sector or private). These independent variables have also been used by Barrientos and Boussofiane (2005), Sathye (2011), and Galagedera and Watson (2015) in their respective studies. The Tobit regression model estimated is presented below;

$urn:x-wiley:14723891:media:pa2806:pa2806-math-0023$

Here, m is the pension fund, n is time, TE is efficiency scores under variable returns of scale, AUM is natural log of asset under management of pension funds, age of funds (AOF) is 1 if more than or equal to 5 years old otherwise 0, Sector is 1 if govt. sector otherwise 0, TOFM (Type of fund manager) is 1 if public sector's fund manager otherwise 0 and $urn:x-wiley:14723891:media:pa2806:pa2806-math-0024$ is error term, and α, β₁, β₂, β₃, and β₄ are parameters.

5 RESULTS AND DISCUSSION

Using the stated inputs and outputs, we have received results for all the pension funds for the financial year 2014–2015. As is clear from Table 4, Column (3), total 16 of the 42 pension funds operated efficiently in the year 2014–2015. The result shows that the average efficiency score for pension funds was 0.8062 and there was significant variation in the efficiency scores of pension funds as the standard deviation was 21.16%. The highest efficiency score was 1, which was achieved by 16 pension funds while the lowest efficiency score was 0.4104 experienced by UTI Retirement Solution Ltd. E2 fund. LIC Pension Fund Ltd. C1, Birla Sunlife C2, HDFC C2, HDFC E1, LIC Pension Fund Ltd. E1, HDFC E2, LIC Pension Fund Ltd. E2, HDFC G1, LIC Pension Fund Ltd. G1, HDFC G2, LIC Pension Fund Ltd. G2, LIC Pension Fund Ltd. APY, SBI Pension Fund Ltd. APY, UTI Retirement Solution Ltd. APY, HDFC A1, ICICI Prudential Pension Funds A1, Kotak Mahindra Pension Fund Ltd. A1, LIC Pension Fund Ltd. A1, Reliance Capital Pension Fund Ltd. A1, SBI Pension Fund Pvt. Ltd. A1 and UTI Retirement Solution Ltd. A1 were not in business in the year 2014–2015.

TABLE 4. Technical efficiency scores and ranks of pension funds (2015–2019)

Sr. no.	Pension fund	2014–2015		2015–2016		2016–2017		2017–2018		2018–2019
Sr. no.	Pension fund	Score	Rank	Score	Rank	Score	Rank	Score	Rank	Score	Rank
1	LIC Pension Fund Ltd. CG	1.0000	1	1.0000	1	1.0000	1	1.0000	1	1.0000	1
2	SBI Pension Fund Pvt. Ltd. CG	1.0000	1	1.0000	1	1.0000	1	1.0000	1	1.0000	1
3	UTI Retirement Solution Ltd. CG	1.0000	1	1.0000	1	1.0000	1	1.0000	1	1.0000	1
4	LIC Pension Fund Ltd. SG	1.0000	1	1.0000	1	1.0000	1	1.0000	1	1.0000	1
5	SBI Pension Fund Pvt. Ltd. SG	1.0000	1	1.0000	1	1.0000	1	1.0000	1	1.0000	1
6	UTI Retirement Solution Ltd. SG	1.0000	1	1.0000	1	1.0000	1	1.0000	1	1.0000	1
7	HDFC C1	—	—	0.9623	2	1.0000	1	0.9296	4	0.9172	7
8	ICICI Prudential Pension Funds C1	0.9950	3	1.0000	1	0.9086	16	1.0000	1	1.0000	1
9	Kotak Mahindra Pension Fund Ltd. C1	1.0000	1	0.9356	4	0.9144	15	0.9796	2	1.0000	1
10	LIC Pension Fund Ltd. C1	—	—	1.0000	1	1.0000	1	1.0000	1	0.8682	10
11	Reliance Capital Pension Fund Ltd. C1	0.9966	2	0.8039	6	0.9551	8	0.8440	8	0.8107	17
12	SBI Pension Fund Pvt. Ltd. C1	1.0000	1	0.8964	5	0.9521	9	0.9592	3	0.9684	4
13	UTI Retirement Solution Ltd. C1	1.0000	1	0.7459	10	0.9582	6	1.0000	1	0.8328	14
14	Birla Sunlife C2	—	—	—	—	—	—	1.0000	1	0.6890	19
15	HDFC C2	—	—	1.0000	1	1.0000	1	1.0000	1	0.8945	8
16	ICICI Prudential Pension Funds C2	1.0000	1	0.9514	3	0.9223	12	1.0000	1	0.9668	5
17	Kotak Mahindra Pension Fund Ltd. C2	0.9856	4	0.6848	13	0.9028	17	0.8276	10	0.8441	12
18	Reliance Capital Pension Fund Ltd. C2	1.0000	1	0.7342	11	0.9567	7	0.8520	7	0.8418	13
19	SBI Pension Fund Pvt. Ltd. C2	0.9649	6	0.7895	7	0.9861	4	0.8581	6	0.8455	11
20	UTI Retirement Solution Ltd. C2	0.9804	5	0.6849	12	0.9984	2	0.9000	5	0.8200	15
21	HDFC E1	—	—	1.0000	1	1.0000	1	1.0000	1	1.0000	1
22	ICICI Prudential Pension Funds E1	1.0000	1	0.3123	27	0.4257	28	0.5388	25	0.5759	23
23	Kotak Mahindra Pension Fund Ltd. E1	0.4997	21	0.2561	32	0.3753	39	0.3922	27	0.4322	38
24	LIC Pension Fund Ltd. E1	—	—	0.2293	37	0.6885	27	0.6221	16	0.8175	16
25	Reliance Capital Pension Fund Ltd. E1	0.6517	20	0.2544	33	0.4221	30	0.3666	28	0.4135	39
26	SBI Pension Fund Pvt. Ltd. E1	0.4558	24	0.2642	29	0.3844	35	0.3056	30	0.3261	41
27	UTI Retirement Solution Ltd. E1	0.7856	8	0.2739	28	0.3777	38	0.5139	26	0.5505	26
28	HDFC E2	—	—	0.1947	39	0.4127	31	0.5577	23	0.6462	20
29	ICICI Prudential Pension Funds E2	0.4496	25	0.2512	35	0.3946	33	0.2904	34	0.2853	46
30	Kotak Mahindra Pension Fund Ltd. E2	0.4649	22	0.2593	31	0.3818	36	0.3156	29	0.3336	40
31	LIC Pension Fund Ltd. E2	—	—	0.2012	38	0.4013	32	0.2933	33	0.2894	45
32	Reliance Capital Pension Fund Ltd. E2	0.4603	23	0.2542	34	0.4246	29	0.2936	32	0.2904	44
33	SBI Pension Fund Pvt. Ltd. E2	0.4490	26	0.2630	30	0.3900	34	0.3028	31	0.2964	43
34	UTI Retirement Solution Ltd. E2	0.4104	27	0.2456	36	0.3795	37	0.2863	35	0.3127	42
35	LIC Pension Fund Ltd. CCG	0.8556	7	0.4541	19	0.7405	26	0.5726	20	0.5742	24
36	SBI Pension Fund Pvt. Ltd. CCG	0.7088	9	0.4115	26	0.7872	25	0.5409	24	0.5475	27
37	HDFC G1	—	—	0.5071	14	0.9151	14	0.7764	11	0.5533	25
38	ICICI Prudential Pension Funds G1	0.6562	19	0.4401	22	0.9218	13	0.5856	18	0.4790	33
39	Kotak Mahindra Pension Fund Ltd. G1	0.7031	10	0.4660	17	0.8653	24	0.7615	12	0.4718	35
40	LIC Pension Fund Ltd. G1	—	—	0.7464	9	0.8854	21	0.8343	9	0.8081	18
41	Reliance Capital Pension Fund Ltd. G1	0.6768	14	0.4409	21	0.8809	22	0.5902	17	0.4652	36
42	SBI Pension Fund Pvt. Ltd. G1	0.6624	17	0.4542	18	0.8883	20	0.5648	21	0.5355	29
43	UTI Retirement Solution Ltd. G1	0.6793	13	0.4305	24	0.9969	3	1.0000	1	0.4598	37
44	HDFC G2	—	—	0.7504	8	0.9679	5	1.0000	1	0.6380	21
45	ICICI Prudential Pension Funds G2	0.6582	18	0.4325	23	0.9271	11	0.5732	19	0.4798	32
46	Kotak Mahindra Pension Fund Ltd. G2	0.6912	11	0.4697	16	0.8920	19	0.7443	14	0.4779	34
47	LIC Pension Fund Ltd. G2	—	—	1.0000	1	1.0000	1	1.0000	1	0.9396	6
48	Reliance Capital Pension Fund Ltd. G2	0.6686	16	0.4410	20	0.8983	18	0.5604	22	0.4949	30
49	SBI Pension Fund Pvt. Ltd. G2	0.6712	15	0.4907	15	0.8749	23	0.7352	15	0.5469	28
50	UTI Retirement Solution Ltd. G2	0.6802	12	0.4223	25	0.9521	10	0.7565	13	0.4876	31
51	Kotak Mahindra Pension Fund Ltd. NPS lite	1.0000	1	1.0000	1	1.0000	1	1.0000	1	1.0000	1
52	LIC Pension Fund Ltd. NPS lite	1.0000	1	1.0000	1	1.0000	1	1.0000	1	1.0000	1
53	SBI Pension Fund Pvt. Ltd. NPS lite	1.0000	1	1.0000	1	1.0000	1	1.0000	1	1.0000	1
54	UTI Retirement Solution Ltd. NPS lite	1.0000	1	1.0000	1	1.0000	1	1.0000	1	1.0000	1
55	LIC Pension Fund Ltd. APY	—	—	—	—	—	—	1.0000	1	1.0000	1
56	SBI Pension Fund Pvt. Ltd. APY	—	—	—	—	—	—	1.0000	1	1.0000	1
57	UTI Retirement Solution Ltd. APY	—	—	—	—	—	—	1.0000	1	1.0000	1
58	HDFC A1	—	—	—	—	—	—	—	—	1.0000	1
59	ICICI Prudential Pension Funds A1	—	—	—	—	—	—	—	—	1.0000	1
60	Kotak Mahindra Pension Fund Ltd. A1	—	—	—	—	—	—	—	—	1.0000	1
61	LIC Pension Fund Ltd. A1	—	—	—	—	—	—	—	—	0.6147	22
62	Reliance Capital Pension Fund Ltd. A1	—	—	—	—	—	—	—	—	0.9960	2
63	SBI Pension Fund Pvt. Ltd. A1	—	—	—	—	—	—	—	—	0.8784	9
64	UTI Retirement Solution Ltd. A1	—	—	—	—	—	—	—	—	0.9843	3
Average		0.8062	—	0.6378	—	0.8171	—	0.7689	—	0.7391	—
SD		0.2116	—	0.3054	—	0.2384	—	0.2571	—	0.2541	—
Minimum		0.4104	—	0.1947	—	0.3753	—	0.2863	—	0.2853	—
Maximum		1.0000	—	1.0000	—	1.0000	—	1.0000	—	1.0000	—
Number of efficient funds		16	—	15	—	15	—	23	—	19	—
Number of inefficient funds		26	—	38	—	38	—	34	—	45	—

Column (5) of Table 4 shows that in the financial year 2015–2016, the number of pension funds increased from 42 to 53 but the number of efficient funds decreased to 15. Six previously efficient funds, namely, ICICI Prudential Pension Funds E1, SBI Pension Fund Pvt. Ltd. C1, UTI Retirement Solution Ltd. C1, Kotak Mahindra Pension Fund Ltd. C1, ICICI Prudential Pension Funds C2, and Reliance Capital Pension Fund Ltd. C2 could not remain efficient in the year 2015–2016. Six new funds, namely, HDFC E1, LIC Pension Fund Ltd. C1, ICICI Prudential Pension Fund C1, HDFC C2, and LIC Pension Fund Ltd. G2 joined the efficient frontier. The lowest efficiency of 0.1947 was experienced by the pension fund HDFC E2. The average efficiency of pension funds decreased to 0.6378 and high variation of 30.54% was observed in the efficiency scores of the pension funds. Birla Sunlife C2, LIC Pension Fund Ltd. APY, SBI Pension Fund Ltd. APY, UTI Retirement Solution Ltd. APY, HDFC A1, ICICI Prudential Pension Funds A1, Kotak Mahindra Pension Fund Ltd. A1, LIC Pension Fund Ltd. A1, Reliance Capital Pension Fund Ltd. A1, SBI Pension Fund Pvt. Ltd. A1 and UTI Retirement Solution Ltd. A1 were not in business in the year 2015–2016.

It is clear from Column (7) of Table 4 that number of efficient funds remained the same, that is, 15 in the financial year 2016–2017 as was in the previous year. ICICI Prudential Pension Fund C1 converted into the inefficient fund while HDFC C1 fund converted into an efficient one. Rest all efficient pension funds were the same as in the previous year. The average efficiency of the pension funds increased to 0.8171 in the year 2016–2017. This was the highest average efficiency score achieved by the pension funds during our study period. Efficiency variation of 23.84% was observed in Pension funds during this year. Minimum efficiency of 0.3753 was observed in Kotak Mahindra Pension Fund Ltd. Birla Sunlife C2, LIC Pension Fund Ltd. APY, SBI Pension Fund Ltd. APY, UTI Retirement Solution Ltd. APY, HDFC A1, ICICI Prudential Pension Funds A1, Kotak Mahindra Pension Fund Ltd. A1, LIC Pension Fund Ltd. A1, Reliance Capital Pension Fund Ltd. A1, SBI Pension Fund Pvt. Ltd. A1 and UTI Retirement Solution Ltd. A1 were not in business in the year 2016–2017.

As is evident from Column (9) of Table 4, the number of total pension funds increased from 53 to 57 in the financial year 2017–2018. The number of efficient funds increased from 15 to 23 in this year. Besides the old ones, the new pension funds that entered the efficient frontier include; UTI Retirement Solution Ltd. C1, ICICI Prudential Pension Funds C1, ICICI Prudential Pension Funds C2, Birla Sunlife C2, UTI Retirement Solution Ltd. G1, HDFC G2, SBI Pension Fund Pvt. Ltd. APY, UTI Retirement Solution Ltd. APY and LIC Pension Fund Ltd. APY whereas HDFC C1 fund left the efficient frontier. The average efficiency of the pension funds decreased to 0.7689 in 2017–2018 as compared to the previous year. A significant variation of 25.71% was observed in the efficiency performance of the pension funds during this year. The lowest efficiency of 0.2863 was observed in the UTI Retirement Solution Ltd. E2 fund. HDFC A1, ICICI Prudential Pension Funds A1, Kotak Mahindra Pension Fund Ltd. A1, LIC Pension Fund Ltd. A1, Reliance Capital Pension Fund Ltd. A1, SBI Pension Fund Pvt. Ltd. A1 and UTI Retirement Solution Ltd. A1 were not in business in the year 2017–2018.

It is evident from Column (11) of Table 4 that total number of pension funds increased to 64 in the financial year 2018–2019, which was the highest number during our study period. The number of efficient funds decreased to 19 in the year 2018–2019. Efficient funds like UTI Retirement Solution Ltd. C1, LIC Pension Fund Ltd. C1, ICICI Prudential Pension Funds C2, HDFC C2, Birla Sunlife C2, UTI Retirement Solution Ltd. G1, LIC Pension Fund Ltd. G2 and HDFC G2 converted into inefficient funds during this year whereas only one fund “Kotak Mahindra Pension Fund Ltd. C1” converted into efficient fund. Average efficiency slightly reduced to 0.7391 in the year 2018–2019. A significant variation of 25.41% was observed in the efficiency levels of the pension funds. The lowest efficiency of 0.2853 was observed in ICICI Prudential Pension Fund E2.

Figure 1 shows that the average efficiency score of Indian Pension Funds drastically reduces in the year 2016 and recoups by 2017 then again decreases in the years 2018 and 2019.

Details are in the caption following the image — **FIGURE 1**
Open in figure viewer PowerPoint

Trend of average efficiency scores of Indian pension funds (2015–2019).
*Source*: Author's calculation

Figure 2 shows approximate uniformity in the efficiency performance of all the pension fund managers except Birla Sunlife. Birla Sunlife achieved higher average efficiency score because it offered only one pension scheme during the study period.

Figure 3 shows variation in the average efficiency performance of different pension schemes. The minimum efficiency was achieved by the Equity Tier II scheme whereas the highest efficiency was achieved by NPS lite, CG, SG, and APY schemes. During our study period, the average efficiency of the Indian pension sector has been found 0.7538 and this sector experienced overall stability in the efficiency performance, about 40% of pension funds operated efficiently. Drazenovic et al. (2019) in their study on Croatian pension funds for the period 2015–2019 found the average efficiency of 0.987. Whereas, Sathye (2011) in his study on Australian pension funds found that the pensions had low efficiencies over the period considered (2005–2009), with mean technical efficiency of 40%, and only 15%–20% of DMUs being technically efficient.

Table 5 shows the relative mean slacks (Murthi et al., 1997). It is a ratio of absolute average slack of an input all over funds to the average value of input all over funds. Relative mean slacks identify the marginal effect of inputs on return that are used inefficiently by fund managers. It is evident from the above table that relative mean slacks in the year 2017–2018 were the least, which brought 40% of the pension funds to the efficient frontier. That was the highest number during our study period. In the year 2016–2017 and 2018–2019, the relative average slack in expense ratio was zero; however, larger slacks were observed in the standard deviation. Similarly, in the year 2015–2016, the relative average slack in expense ratio was zero; however, larger slack was observed in the beta. In the year, 2014–2015, relative mean slack was observed in the expense ratio but 38% of the total pension funds were found efficient. This was higher than in the years 2015–2016 (28%), 2016–2017 (28%), and 2018–2019 (30%) in which zero relative mean slack was observed in the expense ratio. This confirms to “mean-variance efficiency hypothesis” (Sengupta, 2003; Sengupta & Zohar, 2001). As per this hypothesis, inefficiency of the funds is because of the associated risk and not the associated expenses.

TABLE 5. Relative mean slacks (2015–2019)

Year	Standard deviation	Beta	Expense ratio	% age of efficient funds
2014–2015	0.1523	0.1651	0.3630	38
2015–2016	0.0119	0.4237	0	28
2016–2017	0.1990	0.0354	0	28
2017–2018	0.0441	−1.8903	0	40
2018–2019	0.4896	0.3140	0	30

Tobit regression is applied using the AER software, R version 3.6. Table 6 reports the results obtained by the Tobit regression. The coefficients are interpreted to analyze the relationship between efficiency changes and independent variables. The results reveal that two of the four explanatory variables are statistically significant. The results given in Table 6 shows that the estimated coefficient for the AUM of the pension funds is positive and statistically significant at 5% level of significance. This indicates that large-sized pension funds perform better and positively affect the efficiency. Barrientos and Boussofiane (2005), Sathye (2011), and Galagedera and Watson (2015) also found that fund size positively affect the performance of the pension funds. The estimated coefficient for the sector type of the pension funds is also positive and statistically significant. This indicates that pension funds for CG and SG employees perform better than the private sector pension funds. Galagedera and Watson (2015) also found that public sector funds perform better than retail and industry funds. All other variables are not found to have any statistically significant effect on the efficiency performance of the pension funds.

TABLE 6. Coefficient estimates from Tobit regression and pooled regression using fund characteristics: Dependent variable—Efficiency scores under VRS (2015–2019) observations = 269

Variables	Coefficients	Std. error	z-value	t-value
Constant	0.763550	0.042212	18.089	17.920
AUM	0.014779	0.007446	1.985	1.966
Age of Fund	−0.138422	0.035947	−3.851	−3.815
Sector type	0.228420	0.062479	3.656	3.622
Fund manager type	−0.002542	0.032686	−0.078	−0.077
R²	—	—	—	0.1691
Wald statistic	54.73	—		—

Abbreviations: AUM, asset under management; VRS, variable return to scale.

6 CONCLUSION

The study has estimated the efficiency of Indian pension funds for the years 2015–2019. BCC model of DEA was used to estimate efficiency estimates. This study has found that the average efficiency of the Indian pension sector was 0.7538 and this sector experienced overall stability in the efficiency performance with only a slight variation of 7.18% in the efficiency scores during the study period. Indian pension sector achieved the maximum efficiency of 0.8171 in the financial year 2016–2017 and minimum efficiency of 0.6378 in the financial year 2015–2016. Overall, the efficiency of Indian pension funds is found to be high. Whereas, pension funds of some countries like Chile and Australia had low-efficiency scores. The pension funds of other countries like Croatia had higher efficiency scores as compared to India. Indian pension sector has expanded during our study period. The numbers of pension funds increased from 42 in the year 2014–2015 to 64 in the year 2018–2019. The number of efficient pension funds increased from 16 to 19 over the years which were 40% of the total funds. Whereas, only 15%–20% of Australian pension funds were found efficient by Sathye (2011) during the period 2005–2009. About 62% of the pension funds were inefficient in the financial year 2014–2015 and this share increased to 70% in the financial year 2018–2019. SBI Pension Fund Pvt. Ltd. CG, UTI Retirement Solution Ltd. CG, LIC Pension Fund Ltd. CG, SBI Pension Fund Pvt. Ltd. SG, UTI Retirement Solution Ltd. SG, LIC Pension Fund Ltd. SG, SBI Pension Fund Pvt. Ltd. NPS lite, UTI Retirement Solution Ltd. NPS lite, LIC Pension Fund Ltd. NPS lite, and Kotak Mahindra Pension Fund Ltd. NPS lite were found efficient throughout the study period.

Then relative mean slacks were calculated to identify the marginal effect of inputs on return. The minimum slack was found in the input “expense ratio.” This shows that risk associated with investments is the cause of inefficiency in the Indian pension funds. This confirms to “mean-variance efficiency hypothesis” (Sengupta, 2003; Sengupta & Zohar, 2001). The efficiency scores were then related to explanatory variables like fund size (AUM), sector, types of fund manager, and age of fund. The relationship was examined using the Tobit regression model. It was found that fund size of pension funds and government sector pension funds are found to have positively associated with technical efficiency. This confirms to findings of Barrientos and Boussofiane (2005), Sathye (2011), and Galagedera and Watson (2015). Whereas, type of fund manager and AOF were found to have nonsignificant association with efficiency of the pension fund.

7 IMPLICATIONS AND LIMITATIONS OF THE STUDY

There are many small pension funds operating in the Indian pension sector. Our study like Sathye (2011) also indicates that economies of scale can be achieved through rationalization of the sector. It is recommended that mergers & acquisitions be encouraged to enhance the efficiency of this sector, which would ultimately benefit the pension subscribers. This is the first study that has assessed the efficiency of the India pension funds and explored the drivers of efficiency. This study will be helpful for the Indian pension fund subscribers in taking informed decisions before making any investment. Our study provides information to the inefficient pension funds on their shortcomings in every aspect of input excess and output shortfall. The pension funds that scored low in efficiency score should seal the leakages regarding commission as well as distribution channel to reduce/stop adverse performance. Based on this understanding, managers can formulate and implement to convert the fund into the efficient one. This study may also be useful for the policymakers, that is, PFRDA of India in formulating future policies as an awareness of the determinants of cost efficiency may assist in formulating policies related to distribution channels, the commission paid, and working of channel members so that sustainability and efficiency of pension fund managers are secured. Due to unavailability of the required data, efficiency performance of the pension fund managers could not be analyzed.

ACKNOWLEDGMENTS

The author is grateful to the anonymous referees of the journal for their extremely useful suggestions to improve the quality of the article.

CONFLICT OF INTEREST

The authors declare no potential conflict of interest.

Biography

Shoaib Alam Siddiqui is an Assistant Professor at the Joseph School of Business Studies, Sam Higginbottom University of Agriculture, Technology and Sciences, Allahabad, India. He holds his Ph.D. in Business Studies. His research interests include Financial Services, Health Economics and Applied Econometrics. He does research on efficiency, productivity, market structure and development economics with special focus on life insurance, health insurance and pension. He has published in various refereed journals such as Managerial and Decision Economics, Global Business Review, Indian Journal of Economics and Development, The Indian Journal of Commerce and Journal of New Business Ventures. He has contributed various chapters in the edited books. He has over 20 years of corporate-academia experience and worked with LIC of India, TATA-AIG Life Insurance Company Limited and Reliance Life Insurance Company Limited in the past.

Open Research

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.

REFERENCES

Andreu, L., Sarto, J. L. & Vicente, L. (2014). Efficiency of the strategic style of pension funds: an application of the variants of the slacks-based measure in DEA. The Journal of Operational Research Society, 65(2), 1886–1895.
10.1057/jors.2013.74
Web of Science® Google Scholar
Banker, R., Charnes, A., & Cooper, W. W. (1984). Some models for estimating technical and scale efficiency. Management Science, 30(9), 1078–1092.
10.1287/mnsc.30.9.1078
Web of Science® Google Scholar
Barrientos, A., & Boussofiane, A. (2005). How efficient are pension fund managers in Chile? Journal of the Contemporary Economics, 9(2), 289–311.
10.1590/S1415-98482005000200003
Google Scholar
Barros, C., Ferro, G. & Romero, C. (2008). Technical efficiency and heterogeneity of Argentina pension funds (working papers 2008/29, Department of Economics at the School of Economics and Management [ISEG], Technical University of Lisbon).
Google Scholar
Barros, C. P., & Garcia, M. T. M. (2006). Performance evaluation of pension funds management companies with data envelopment analysis. Risk Management and Insurance Review, 9(2), 165–188.
10.1111/j.1540-6296.2006.00092.x
Google Scholar
Basso, A., & Funari, S. (2001). A data envelopment analysis approach to measure the mutual fund performance. European Journal of Operational Research, 135, 477–492.
10.1016/S0377-2217(00)00311-8
Web of Science® Google Scholar
Bikker, J., & Dreu, J. D. (2006). Operating costs of pension funds: The impact of scale, governance and plan design. Journal of Pension Economics and Finance, 8(1), 63–89.
10.1017/S1474747207002995
Web of Science® Google Scholar
Braberman, D. G., Chisari, O. O., Quesada, L., & Rossi, M. A. (1999). Funcion de costos en la industria de lasadministradoras de fondos de jubilaciones y pensiones en la Argentina: Un analisis de lasmodificaciones regulatorias. Revista de Análisis Económico, 14(1), 121–142.
Google Scholar
Charnes, A., Cooper, W. W., & Rhodes, E. (1978). Measuring the efficiency of decision making units. European Journal of Operational Research, 2(6), 429–444.
10.1016/0377-2217(78)90138-8
Web of Science® Google Scholar
Choi, Y., & Murthi, B. (2001). Relative performance evaluation of mutual fund: A non-parametric approach. Journal of Business Finance and Accounting, 28(7/8), 853–876.
10.1111/1468-5957.00396
Google Scholar
Daraio, C., & Simar, L. (2006). A robust non-parametric approach to evaluate and explain the performance of mutual funds. European Journal of Operational Research, 175, 516–542.
10.1016/j.ejor.2005.06.010
Web of Science® Google Scholar
Drazenovic, B. O., Hodzic, S., & Maradin, D. (2019). The efficiency of mandatory pension funds: Case of Croatia. South East European Journal of Economics and Business, 14(2), 82–94.
10.2478/jeb-2019-0015
Web of Science® Google Scholar
Galagedera, D. U. A., & Watson, J. (2015). Benchmarking superannuation funds based on relative performance. Applied Economics, 47(28), 2959–2973.
10.1080/00036846.2015.1011315
Web of Science® Google Scholar
Hollingsworth, B., & Smith, P. (2003). The use of ratios in data envelopment analysis. Applied Economics Letters, 10(11), 733–735.
10.1080/1350485032000133381
Web of Science® Google Scholar
Lin, S. W., Lu, W. N., & Lin, F. (2020). Entrusting decisions to the public service pension fund: An integrated predictive model with additive network DEA approach. Journal of operational research society, 72, 1015–1032. https://doi.org/10.1080/01605682.2020.1718011
10.1080/01605682.2020.1718011
Web of Science® Google Scholar
Mogha, S. K. (2020). Sensitivity in efficiency and super efficiency evaluation: Case of a private educational institution. International Journal of Operational Research, 1, 1. https://doi.org/10.1504/IJOR.2020.10035261
10.1504/IJOR.2020.10035261
Google Scholar
Mogha, S. K., Yadav, S. P., & Singh, S. P. (2014). Estimating technical and scale efficiencies of private hospitals using a non-parametric approach: Case of India. International Journal of Operational Research, 20(1), 21–40.
10.1504/IJOR.2014.060515
Google Scholar
Mogha, S. K., Yadav, S. P., & Singh, S. P. (2016). Estimating technical efficiency of public sector hospitals of Uttarakhand (India). International Journal of Operational Research, 25(3), 371–399.
10.1504/IJOR.2016.074759
Google Scholar
Morey, M. R., & Morey, R. C. (1999). Mutual fund performance appraisals: A multi-horizon perceptive with endogenous benchmarking. Omega, 27, 241–258.
10.1016/S0305-0483(98)00043-7
Web of Science® Google Scholar
Murthi, B. P. S., Choi, Y. K., & Desai, P. (1997). Efficiency of mutual funds and portfolio performance measurement: A non-parametric approach. European Journal of Operational Research, 98, 408–418.
10.1016/S0377-2217(96)00356-6
Web of Science® Google Scholar
New Pension Scheme Trust (2018–2019). Annual Reports, 2014–15, 2015–16, 2016–17, 2017–18 and 2018–2019. www.npstrust.org.in
Google Scholar
Pastor, J. T., & Lovell, C. A. K. (1995). Units invariant and translation invariant DEA models. Operations Research Letters, 1, 147–151.
Google Scholar
Russell, R. (1988). On the axiomatic approach to the measurement of technical efficiency. In Measurement in economics: Theory and applications of economic indices. Physica-Verlag.
10.1007/978-3-642-52481-3_18
Google Scholar
Sathye, M. (2011). The impact of financial crisis on the efficiency of superannuation funds: Evidence for Australia. Journal of Law and Financial Management, 10(2), 16–27.
Google Scholar
Sedzro, K. & Sardano, D. (1999). Mutual fund performance evaluation using data envelopment analysis, Working paper, School of Business, University of Quebec.
Google Scholar
Sengupta, J. (2003). Efficiency tests for mutual fund portfolios. Applied Financial Economics, 13, 869–876.
10.1080/09603100210161992
Google Scholar
Sengupta, J., & Zohar, T. (2001). Non-parametric analysis of portfolio efficiency. Applied Economics Letters, 8, 249–252.
10.1080/135048501750103999
Web of Science® Google Scholar
Sharma, G., & Sharma, V. (2018). Performance evaluation of equity mutual funds using DEA. International Journal of Financial Services Management, 9(1), 1–13.
10.1504/IJFSM.2018.089915
Google Scholar
Sharpe, W. (1966). Mutual fund performance. Journal of Business, 34, 119–138.
10.1086/294846
Web of Science® Google Scholar

Volume22, IssueS1

December 2022

e2806

Efficiency evaluation of the pension funds: Evidence from India

Abstract

1 INTRODUCTION

2 REVIEW OF LITERATURE

3 OBJECTIVES

4 DATA AND METHODOLOGY

4.1 CCR and BCC input-oriented models

4.2 Specification of the regression model

5 RESULTS AND DISCUSSION

6 CONCLUSION

7 IMPLICATIONS AND LIMITATIONS OF THE STUDY

ACKNOWLEDGMENTS

CONFLICT OF INTEREST

Biography

Open Research

DATA AVAILABILITY STATEMENT

REFERENCES

Figures

References

Information

About Wiley Online Library

Help & Support

Opportunities

Connect with Wiley

Efficiency evaluation of the pension funds: Evidence from India

Abstract

1 INTRODUCTION

2 REVIEW OF LITERATURE

3 OBJECTIVES

4 DATA AND METHODOLOGY

4.1 CCR and BCC input-oriented models

4.2 Specification of the regression model

5 RESULTS AND DISCUSSION

6 CONCLUSION

7 IMPLICATIONS AND LIMITATIONS OF THE STUDY

ACKNOWLEDGMENTS

CONFLICT OF INTEREST

Biography

Open Research

DATA AVAILABILITY STATEMENT

REFERENCES

Figures

References

Related

Information