FEEM PPT Template

1
Teoria delle aste e applicazioni in finanza
Lezione 8Privatizzazioni Teoria e Fatti
Stilizzati

Docente Bernardo Bortolotti
2
What is privatization?

• Definition
• a transfer of ownership or voting rights in a
  State-owned enterprise from the State to the
  private sector. The State includes the central or
  local government, ministries, bodies of the
  public administration, and companies where the
  central or local government acts as shareholder.
  The private sector comprises private individuals
  and economic entities with private shareholders.

3
State vs private ownership theory

• A classic starting point the Fundamental
  Theorems of Welfare Economics
• A competitive equilibrium is Pareto-efficient
• Any Pareto efficient allocation can be supported
  by a competitive equilibrium
• Under strong assumptions, when markets are
  competitive state ownership of firms cannot be
  justified on efficiency grounds

4
State vs private ownership theory

• Market failures (info, monopoly, externalities,
  public goods, etc.) provide a rationale for
  government intervention in the economy
• Such a rationale does not support state
  ownership the government can provide goods and
  services without producing them
• Neutrality theorems (Sappington and Stiglitz
  1987, Shapiro and Willig 1990)

5
Sappington and Stiglitz (1987)

• Fundamental Privatization Theorem
• Suppose government pursues these objectives (a)
  economic efficiency (b) equity (c) rent
  extraction. Then the government can design an
  auction scheme to meet perfectly these
  objectives, and public production cannot improve
  private production

6
Sappington and Stiglitz (1987)

• Proof (sketch)
• Gov has a social valuation of the good (includes
  equity).
• Production by (risk-neutral) firms.
• Public procurement mechanism the winning firm
  will receive a payment to social valuation.
• The most efficient firm will win the contract
• If auction competitive, complete rent extraction

7
Politicians in firms Vishny and Shleifer (1994)

• Assumptions
• Politicians control SOEs to achieve political
  objectives (excess labor, rents, etc.)
• Three players the treasury (non strategic) the
  politician, the manager
• Separation of cash flow and control rights
• Commercialization gov has cash flow rights, but
  no control rights
• Regulation manager has ownership rights but
  govt has control rights

8
The incomplete contracts approach

• Neutrality results hinge upon availability of
  complete contingent long-term contracts. When
  contracts are incomplete, ownership matters
  (Williamson 1985, Grossman and Hart 1986)
• Suppose that surplus depends upon on some
  non-contractible, relation specific investment.
  The division of these quasi-rents will be
  determined by the ex post bargaining power of the
  parties, which in turn depends on the governance
  structure
• Privatization matter because makes the private
  owner is the residual claimant of the surplus.

9
Sketch of Schmidt (1996)

• Monopolistic firm produces y of a public good,
  generating social benefit b(y)
• Costs are c(y,?) where ? is a state variable
  (high, low) which is private information of the
  owner of the firm
• Manager makes a personal, unobservable
  non-monetary investment, e, to reduce future
  costs
• q(e) is the probability of the low cost state ?

10
Sketch of Schmidt (1996)

• Time line

11
Sketch of Schmidt (1996)

• Nationalization
• Suppose that contingent contracts are not
  feasible (wages and subsides are fixed)
• The government then chooses the ex post efficient
  production level y
• Given payments are fixed, managers sets e0,
  government will offer w0
• Lack of commitment of the parts of government
  induces efficient production level, but
  productive inefficiency (costs are high because
  manager does not exert effort)

12
Sketch of Schmidt (1996)

• Privatization
• The government does not know ?, nor the effort
  taken by the manager.
• Standard mechanism design problem (Baron and
  Myerson 1982) to induce the efficient level of
  production, the govt must provide an
  informational rent
• The optimal regulation scheme distorts
  production below the efficient level if costs are
  high. The manager has now incentives to invest in
  cost reduction to get the rent

13
Sketch of Schmidt (1996)

• Key results
• Privatization entails a trade-off between
  allocative efficiency (production levels) and
  productive efficiency (investment levels)
• Normative implication privatization superior
  when social benefit is small
• However, under these circumstances privatization
  is superior even when the govt is fully
  benevolent and rational player
• Neutrality fails
• Related papers Laffont and Tirole (1991),
  Shapiro and Willig (1990)

14
The political economy of privatization

• Previous research assumed governments
  preferences as exogenously given (i.e. benevolent
  (Schmidt 1996) vs self-interested (Shleifer and
  Vishny 1994))
• More recent research in political economy focused
  on voters preferences affecting privatization
  outcomes and choices
• Biais and Perotti (2002), Roland and Verdier
  (1994), Schmidt (2000)

15
Mass privatization Roland and Verdier (1994)

• Context massive transfer of ownership rights by
  free (or quasi free) distribution)
• Irreversibility give the domestic population a
  stake in the success of privatization
• Fast process exploit political windows of
  opportunity
• Czech Rep, Russia (insiders)

16
Mass privatization in CEE and FSU countries
17
Mass privatization Roland and Verdier (1994)

• Hyp Govt maximized the utility of the average
  citizen
• Stage 1
• all firms are offered to private investors
  (privatized)
• Investors decide to enter and restructure, with a
  fixed cost k
• State 2
• Govt observes the level of privatization, and may
  re-nationalize if the level is lower than
  expected
• Production takes place and unemployment occurs

18
Mass privatization Roland and Verdier (1994)

• Trade-off a higher of privatizations increases
  the probability of unemployment, but increases
  the income of the employed as redundant workers
  are no longer subsidized with taxation
• At equilibrium, marginal costs will equate
  benefits
• Unique equilibrium with full, partial
  privatization, or re-nationalization depending on
  entry costs k

19
Mass privatization Roland and Verdier (1994)

• New hyp Positive externalities increasing with
  private sector development (spillovers and
  complementarities in terms of product quality,
  knowledge, RD, etc.)
• Privatization (if sufficiently large) may
  decrease the probability of unemployment
• Multiple equilibria irreversibility if
  investment above a critical level. Otherwise,
  below that level, risk of policy reversal
• Mass privatization free distribution of shares
  may couterbalance the negative effects on
  unemployment since citizens receive more
  dividends. Full priva becomes the only equilibrium

20
Credible privatization Perotti (1995)

• Motivation privatization can be designed to
  signal commitment to market oriented policies
• Hyp
• two types of governments populist and committed
• As in Biais and Perotti, if SOE privatized,
  manager exerts efffort
• Stage 1
• Govt. decides to sell
• Production takes place
• Govt decides to interfere (regulation,
  nationalization)
• Profits distrubuted
• Stage 2
• Govt. sells residual stake
• Production
• Interference

21
Credible privatization Perotti (1995)

• Results when there is no uncertainty over govt.
  preferences, govt maximize utility with a
  complete immediate sale.
• Obviuosly, the firm net worth (and revenues) will
  be lower when government is populist as
  interference will adversely affect managers
  incentives
• Assume private information. Now populist has
  incentives to mimick the committed government
• Results
• There exists a separating equilibrium where the
  committed government sells a small initial stake
  as a (credible) commitment device.
• The populist government will sell instead its
  stake upfront

22
Credible privatization Perotti (1995)

• This signalling equilibrium hinges upon the
  (problematic) assumption that the first tranche
  involves the transfer of control
• If a large stake is needed to relinquish control,
  committed government need an additional
  instrument to signal commitment
• Undepricing of shares selling the first tranche
  at discount, committed gov. signals patience to
  garner privatization benefits overtime
• Populist government will sell entire capital at
  market price (UK vs France)

23
Sketch of privatization history

• 1961, Federal Republic of Germany. Adenauer (CDU)
  sells shares in Volkswagen and VEBA. Govt bails
  out
• 1973, Chile. Pinochet attempts to privatize
  companies nationalized under Allende
• 1979-97, UK. Still the most important. SOE value
  added/GDP from 10 to 0.
• 1987-88, Japan. Nippon Telephone and Telegraph
  (NTT) the largest share offering in history
  US40bn
• 1990s, Europe. Italy, Germany, France, Spain
• 1992-1996, Mass privatization in Russia and Czech
  Rep.
• Mid 1990s CEEC and FSU

24
Facts

• From 1977 to 2004, US1.26trn revenues and 4,500
  privatization deals in more than 100 countries
• Global State-owned Enterprise (SOE) value added
  ( GDP) decrease from 9 to 6
• SOE market capitalization US3,310bn
• 2004
• 164 operations (75bn) globally
• 53 of operations and 72 of revenues in EU
• 58 in 2004 privatization revenues with respect
  to 2001-2003 average
• Privatization coming back strong after end of
  century plunge.

25
Global Privatizations, 1977-2004
26
SOE of GDP by stage of national development
27
Country ranking by revenues
28
Country ranking by revenues
29
Revenues distribution by sectors
30
Last major deals
31
What is left to sell in Europe?

• The central governments (or privatization
  agencies) portfolio in listed companies is worth
  230bn
• 75 of this portfolio is directly held 25 is
  held by financial holdings controlled by the
  State (Kfw, CDP, CDC, etc.)
• This portfolio is worth approximately 1/3 of
  total privatization revenues. Privatizing the
  final tranche becomes the issue. Will
  government relinquish control?
• More likely, IPO wave of non-listed State-owned
  firms

32
European governments portfolio
33
What is left to sell in Europe?

• The central governments (or privatization
  agencies) portfolio in listed companies is worth
  230bn
• 75 of this portfolio is directly held 25 is
  held by financial holdings controlled by the
  State (Kfw, CDP, CDC, etc.)
• This portfolio is worth approximately 1/3 of
  total privatization revenues. Privatizing the
  final tranche becomes the issue. Will
  government relinquish control?
• More likely, IPO wave of non-listed State-owned
  firms

34
Ownership and control Two basic questions

• Did governments give up control in privatized
  firms?
• If they did not, does governments reluctance
  affect firm value?

35
Definition

• We define Reluctant Privatization as
• A privatization of a State-owned Enterprise (SOE)
  characterized by the trasfer of ownership rights
  without a (corresponding) transfer of control
  rights

36
Reluctant Privatizations, Italy

• In December 2003, the Italian MEF corporatized
  Cassa Depositi e Prestiti (CDP), a public
  financial institution, and then trasferred stakes
  in Eni, Enel, Poste Italiane, and STM to CDP.
• 30 of CDP was privatized to a consortium of
  Italian banking foundations.
• 13.1bn privatization revenues cleared by
  Eurostat to amortize public debt.

37
Reluctant Privatizations, Italy
38
Reluctant Privatizations, Italy
39
Research Design

• Did governments give up control in privatized
  firms?
• We analyze empirically the evolution of ultimate
  control rights (from end-1996 to 2000) in
  privatized companies as opposed to a control
  sample.
• If they did not, does governments reluctance
  affect firm value?
• We estimate the (adjusted) performance of
  privatized companies as a function of
  (government) control rights.

40
Theoretical Predictions

• The political interference theory
• (Shleifer and Vishny, QJE1994)
• Governments run SOEs to achieve political
  objectives (high employment, high wages, etc.)
  and forgo maximizing profits.
• H0 Government control rights are not negatively
  discounted in market values.

41
Novelty

• Huge literature on performance of privatized
  companies (Megginson and Netter JEL02 survey).
• To our knowledge, our paper is the first to study
  in combination
• ultimate control (La Porta et al JF02, Tian WP02)
• additional control devices (golden shares)
• adjusted performance using a control sample at
  the company level (Dewenter and Malatesta AER01,
  Lopez-de-Silanes QJE96)

42
Data

• 141 companies privatized by public offerings
  before 31/12/1996 in OECD countries
• Complete ownership data for 1996 and 2000
• 104 privatization prospectuses to identify
  statutory constraints and additional control
  devices
• Balance sheet data for the period 1996-2000
• Main sources SDC Platinum, WordScope,
  Datastream, national exchanges, etc.

43
Measuring Control Rights

• We rely on ultimate control (voting) rights
  (weakest link concept, 10 cut off)
• Six types of ultimate owners
• a family or an individual
• the State
• a widely-held financial institution
• a widely-held corporation
• a miscellaneous investor
• a cross-holding
• (A company that does not have a controlling
  shareholder at the 10 cut-off is classified as
  widely-held).

44
Examples Lufthansa AG, 1996
Govt. direct ownership 1.77 Govt ultimate
control rights 50.7
45
Examples SGS-Thomson Microelectronics, 1996
Govt. direct ownership 0 Govt ultimate control
rights 100
46
The control sample

• First-best a) country b) sector and c) 30
  range of market cap
• 68 of cases
• Second-best a) sector and b) 30 range of
  market cap
• 30 of cases
• Third-best case a) country and b) 30 range of
  market cap
• 1 case

47
The largest shareholders in privatized companies
48
Control rights in privatized vs control sample
49
Golden Shares

• Governments may enjoy control power in partially
  or even in fully privatized firms through golden
  share provisions (GSP)
• GSP typically involve
• Special powers to the public shareholder (veto or
  agreement on MA, representatives in BoD, etc.)
• Statutory constraints in the company bylaws to
  curb private ownership rights (ownership limits,
  voting caps, national control provision, etc.)

50
Golden Shares
GSP are widespread in several countries, highly
correlated, and tend to persist overtime. Not
surprisingly, they are typically concentrated in
strategic sectors, such as aerospace and
defence, oil and gas, and utilities. The
rationale is the protection of relevant national
interests (the public counterpart of poison
pills).
51
Golden Share Provisions
52
Main results on control rights (Summary)

• In 2000, the government is the largest ultimate
  shareholder in 30 of privatized companies.
• These companies are much more tightly controlled
  than their private peers.
• In 2000, the government controls directly
  (through stakes) or indirectly (through GSP) 62
  of privatized companies (65 in 1996).
• Thus, governments privatized SOEs but did not
  relinquish control
• ? Reluctant privatization

53
Market valuation

• Does govt. reluctance affect valuation?
• First, we study the evolution of the adjusted
  market valuation (market-to-book ratio) in
  sub-samples of govt vs non govt controlled
  privatized companies.
• Second, we perform panel data estimation of M/B
  as a function of governments (residual) control
  rights.

54
Market valuation (M/B) descriptive analysis
55
Market valuation (M/B) econometric analysis
Govt. controlled firms have higher (adjusted) M/B
than firms privatized more fully. A thorough
empirical test needed. We estimate the following
model
Where STATE is the covariate for government
ultimate control rights, SPECIAL is a GSP dummy,
and a vector of (pair-specific) control
variables
56
OLS results (random effects)
57
Endogeneity of market valuation

• Abnormal market performance typically drives
  privatization, especially when govt. is in
  financial distress.
• 2SLS estimation relaxing the exogeneity of govt.
  stakes.
• Choice of IV Political-institutional variables
  (majoritarian vs consentual pattern of
  democracy), the partisan orientation of
  privatizing governments, public finance
  conditions.
• These IV are valid strongly correlated with
  stakes, but uncorrelated with the error term.

58
2SLS estimates of M/B
59
Summary of results

• Government stakes do matter for valuation. But
  quite surprisingly higher stakes are associated
  with higher valuation differentials.
• Impact not negligible a 10 percent increase in
  voting rights (approx a standard deviation
  change) brings about 0.6 (25 percent) increase in
  adjusted M/B.
• Therefore, fully divested companies appear less
  valuable than companies where government owns
  large stakes.

60
Conclusion

• We find evidence broadly consistent with the
  concept of Reluctant Privatization. Indeed,
  privatizing government do not give up control in
  SOEs.
• Quite surprisingly, government reluctance does
  not attect negatively market valuation.
• The null hypothesis on political interference
  cannot be rejected.
• Interpretation Grabbing hand story very
  relevant at the intial stage of the process. But
  the government as large shareholder in privatized
  firms may also provide benefits such as
  contracts, favourable regulation, and stability

61
Financial performance of privatized firms

• Idiosyncratic factors may affect the risk and
  return characteristics of privatized firms
• Scope for efficiency gains
• Sectors of activity (energy, utilities, TLC,
  banking)
• Large enterprises with a broad clientele
• Manage large infrastructures
• Often regulated and shielded from competition
• Peculiar pricing, sequencing, and ownership
  structures
• Players in MA (bidder and targets)

62
PB Index

• The PB (Privatization Barometer) Index is a
  benchmark for tracking the performance of
  privatized companies.
• It serves primarily as
• Benchmark for portfolio managers and investors
  who invest in privatized companies
• Performance yardstick for governments and
  investment banks floating shares of state-owned
  companies
• Vehicle for attracting attention to
  privatization-related funds in European equity
  markets

63
PB Index

• The PB Index is developed by the Privatization
  Barometer - a web-based info-provider on
  privatization in the EU - with the support of
  Lehman Brothers
• The index is restricted to ordinary shares of
  privatized companies trading in the stock
  exchanges of the European Union, including the
  ten new accession countries.

64
PB Index

• PB Composite Index
• PB Old Europe (EU15)
• PB New Europe (10 new accession countries)
• PB Banking
• PB Industrial
• PB Oil Gas
• PB Utilities
• PB Telecom

65
PB Index Constituents

• The constituents are the stocks of companies
  privatized from January 1977 to date.
• A privatization is defined as a transfer of
  ownership or voting rights from the central or
  local government, or from bodies of the public
  administration, to private investors.
• Eligible securities are also traded shares of
  equity carve-outs from state-owned enterprises
  (Postabank rt or privatized companies (Terna,
  Page Jaunes, etc.)
• Privatization transactions are identified from
  the Global New Issues Database of Thomson-SDC,
  and double-checked from official sources.

66
PB Index Constituents
as of 23/05/05.
67
PB Index Maintenance

• Adjustment for company additions and deletions
  and stock price adjustments due to corporate
  actions and merger and acquisitions (MA)
  activity are made on a quarterly basis.
• In case of MA, the privatized companys share
  price is replaced by the one of the resulting
  newco in case of merger, and by the one of the
  acquiror in case of an acquisition, if
• the acquiror is listed in the same stock market
  where the privatized company was initially
  traded.
• the acquirors mkt cap is not more than double
  the target

68
Calculations

• The index is capitalization-weighted, and
  calculated with the Laspeyres formula, which
  measures price changes against a fixed base
  quantity weight.
• The index is calculated on a pure price basis
  (not total return), and uses daily data and
  closing prices of the stocks at the base date and
  at each date.
• The source for price and quantity data is
  Datastream.

69
Concentration Ratios by Market Cap
as of 23/05/05.
70
PB Index Coverage vs. Benchmark
as of 23/05/05
71
PB Index Performance
72
PB Index Performance
73
PB Index Performance
74
PB Index Performance
75
PB Index Performance
76
PB Index Performance
77
PB Index Returns
78
PB Index Average Excess Returns
NOTE Average excess return indicates the
historic average differential return of the index
to its respective benchmark. All values are
annualized.
79
PB Index Volatility vs Benchmark
NOTE All values are annualized.
80
Risk-adjusted returns Information Ratios
NOTE Information Ratio indicates the historic
average differential return of the index to its
respective benchmark per unit of historic
variability of the differential return.
81
Risk-adjusted returns Sharpe Ratios
NOTE Sharpe Ratio indicates the historic average
differential return of the index over a risk-free
asset (Germany Interbank Rate 3m) per unit of
historic variability of the differential return.
82
PB Index and CAPM
NOTE Estimation is based on daily data.
Sectorial controls included.
83
Conclusion

• Strong overperformance of several PB Indexes
  (Composite ? 6 over the 3 years period)
• Overperformance survives when adjusted for risk
• Incremental performance of PB over well
  diversified porfolios (Jensens ? ? 6, 10 on an
  annual basis for PB Composite and Industrial,
  respectively)
• Open issues
• Research identifying the source of idiosyncratic
  risk for privatized companies
• Does residual ownership matters?