Week 4

Unions emerged with the Industrial Revolution

Nature of factory work created a new kind of labour force (agricultural → industrial)

Unions fiercely opposed by employers and government

1800s – growing middle class sympathy for union demands

British trade unions legalised in 1872 → growth of unions in other industrialising countries

20th C: One long war – fought on battlefield – but also in the factory (production crucial for winning side)

Unions grew during 20th C – peaking 1950s-70s

Does anyone know who Sidney & Beatrice Webb were?

Notes:

The Webbs viewed Trade Unions as democratic organisations, and defended unions as ‘a necessary element in the democratic state’. That is democracy included economic as well as political relations. In their view, industrial democracy meant a ‘system of industrial relations in which arbitrary and irresponsible power’ was ‘replaced by the continual adjustment of interests between the leaders accountable to the workers on one hand and the public servants accountable to the community as a while on the other’.

1. The share of workers who are trade union members or fall under collective agreements has significantly declined in recent decades

2. The tendency towards new, and often precarious, forms of labour

3. The individualisation of employment relationship

-Decline in unionisation/ strikes mirrored by rise in cases taken to industrial tribunal

-Legislation curbing union power mirrored by strengthening of individual rights and firms’ investment in other (direct) forms of voice

-Is individual(ised) conflict the ‘flipside’ of declining union power?

Decline in the UK was at first traced back to Thatcher-era (1980s) legislation, but this is a wider, international trend and clearly there are multiple and varied explanations specific to different contexts.

Ideological: anti-union laws

Privatisation: much easier to unionize government monopolies

Globalisation: increasing competition undermined single-country union power

Deindustrialisation: Shift from heavy manufacturing to services

Technological change: automation – robots don’t join unions

Changing employment patters: More women

Implication: Individualization of employment rights (…)

…MISSING FROM SLIDES

Two Views:

Neoclassical Economics: YES

Labour subject to supply and demand

Freely entered contracts

Market determines wages

Institutional Economics: NO - Labour is Special

Why Employment is Different (Simon, 1951):

Incomplete contracting - Can't specify everything in advance

Information asymmetry - Can't observe effort perfectly

Small number exchange - Not infinite buyers/sellers like commodities

Power imbalance - Inherent conflict of interest between worker and employer

Evidence: Internal Labour Markets

Firms deliberately insulate from external markets

Wages set by internal rules, not market rates

Internal promotion, stable employment

If labour were just a commodity, why create these systems?

Bottom Line:

Employment is a relational contract - an ongoing relationship with terms that can't all be agreed in advance. It's fundamentally different from buying/selling a commodity.

Neoclassical economics ignores much of the complexity of the employment relationship as noise.

(Aims to set out abstract simplified mathematical propositions)

Model’s employment relationship as subject to laws of supply and demand, freely entered into by worker and employer

Ignores central conflict of interest/ power

Institutional economists stepped into this void to consider the (messy) reality of how employers dealt with the employment contract in practice

Simon 1951 & Williamson 1975

Simon is one of the earliest (and certainly most influential) scholar to discuss the distinctive nature of the employment contract.

He notes that employment contracts are very different from ‘sales’ contracts (the type of contract assumed in the pages of an economics textbook where the commodity to be exchanged and its price is completely specified, and both buyer and seller know exactly what they are getting in the exchange).

In contrast, with an employment contract the employee ‘agrees to accept the authority’ of the employer in return for a given wage. There is considerable uncertainty for both sides about the nature of the exchange.

Simon (1951): Economic theory ignores the two most striking things about the employment contract:

1. That it is so different from a ‘sales’ contract

2. How employers actually manage the employment contract

In reality employment relationship typically characterised by:

1. Incomplete contracting

2. Information asymmetry

3. Small number exchange conditions

Notes:

Simon is one of the earliest (and certainly most influential) scholar to discuss the distinctive nature of the employment contract.

He notes that employment contracts are very different from ‘sales’ contracts

A “sales” contract is complete – the seller promises an exact commodity, and the buyer an exact price. Each side knows exactly what they are getting in advance. This is the type of market contract assumed in economics. An example of a labour market that looks close to a sales contract is Uber, or the market for unskilled labourers. But most employment contracts are not like this.

Incomplete contracting – give myself as an example in terms of LSE contract.

Information asymmetry – you know much more about yourself and what you are able to do than your employer. And the reverse, your employer knows much more about the job than you do. Because of labour asymmetry the economist’s assumptions of complete information are not satisfied.

Small number of exchange conditions – Most employees cannot really be regarded as completely homogenous interchangeable units. When you go for a job people are going to want to know a lot more about you. The more senior and more specialist the role the fewer number of exchanges get made, and so the laws of supply and demand do not seem to work so clearly.

In other words, the assumption of lots of buyers and lots of sellers and a homogenous product might be true of some types of labour markets (Uber), but it is not typical.

1. Lots of buyers and lots of sellers

2. A homogenous product

3. Perfect information

4. Complete contracts

Might be true(ish) of some types of labour markets, but it is certainly not typical

Can you think of examples of labour markets that are like this?

1. Gig/Platform Work

Uber/Deliveroo drivers - many buyers/sellers, simple task, app provides perfect information, one-off transactions

2. Agricultural Seasonal Work

Fruit picking, harvest labour - many workers/farms, homogenous simple tasks, short-term spot contracts

3. Day Labourers

Construction/manual labour hired daily - many available workers, basic standardised tasks, pay agreed upfront

4. Online Freelance Platforms (Simple Tasks)

Fiverr/Upwork for data entry, basic design - global marketplace, clear deliverables, one-off contracts

1. Incomplete contract terms

2. Time component – fixed or open-ended?

3. Asymmetric information

4. Measurement problems and what constitutes performance

5. Small numbers contracting

6. Voluntary assignment of authority (the “area of acceptance” – Simon, 1951)

1. Spot contracting: (A series of one-shot deals) a market type contract where the employer and employee sustain no long-term commitment. A simple example might be a casually employed gardener or window cleaner. Or Gig economy work (Uber). Day hiring of labourers in construction. There are advantages and disadvantages to this type of contract – For er: labour does not become a fixed cost. But hard for the firm to develop specific skills. For ee: they get flexibility and freedom, but also the risk.

2. Contingent claims: (Negotiating detailed terms and conditions beforehand (short-term written temporary contracts)) try to write a comprehensive contract itemising every employee task and setting a price for it. This can become difficult when events not originally covered come up, and then haggling (ie transaction costs) re appear. Advantage for both sides is a reduction in uncertainty. Disadvantage on both sides is: transaction costs when new situations occur.

3. Authority relationship (or a relational contract): (A contract where there is an agreement to have an (employment) relationship but without being able to agree all terms in advance) W leans heavily on D&P’s idea of an internal labour market.

For example:

Uber, fruit pickers, day labourer's

LSE guest teachers

Most of the jobs the students will have

All these different types of contract have advantages and disadvantages. From an economic perspective, in a competitive market, firms that are best able to use contracts that maximise the benefits will prosper

In practice (in contrast to presumption in neoclassical economics) many employers construct systems that are insulated from external labour markets

-Pricing and allocation of labour governed by a set of administrative rules and procedures

-Prescribed points of entry and exit

-Other jobs filled by internal promotion or transfers

-Wages set according to an internal matrix with differentials prescribed by custom and practice (relatively insulated from market)

-On-the-job firm specific training

-Set promotion criteria

-Relatively stable employment / reduced job turnover

-Commitment and identity effects

Notes:

D&P economists who noted that while the presumption in neoclassical economics is that the most efficient system is a highly competitive one, that in practice many employers choose to construct employment systems that are in fact rather insulated from the external labour market.

They called these internal labour markets.

Their research focussed primarily on blue collar employment, but you can see this type of structure in many other industries and types of occupation.

They argue that ILMs appear to be generated by a series of factors not envisioned in conventional economic theory:

Skill specificity. Where firms need to have employees that have very firm specific skills, they will have to train those employees, and thus will try to find ways to reduce labour turn over, and one way to do that is to construct an internal labour market.

On the job training. This is related to the above. Even if the skills needed by the firm are specific, so there is a reduced incentive for individuals to pay for it, in many cases it is not just a matter of who pays for the training (as in the above) but how they are trained. In many cases on the job training will be essential, and so again firms need to construct their own labour markets.

Customary law. They also note that the impact of social norms is generally ignored by economists. And one of the reasons for ILMs is that it means that firms can establish a set of rules, and this creates both stability and a sense of fairness. Both of which are enormously valued by employees.

In sum in an ILM the central characteristic is that rather than the pricing and allocation of labour being done competitively on the open market, what employers do is to decide the price and allocation of labour according to a set of rules and procedures.

They argue that ILMs are generally popular with employees, but can also add value in certain circumstances to employers.

The next slide looks at the factors that tend to make ILMs relatively more attractive to employers.

1. Skill specificity. Where firms require firm specific skills, they will try to find ways to reduce labour turn over, and one way to do that is to construct an internal labour market.

2. On the job training. In many cases on the job training will be essential and so again firms need to construct their own labour markets.

3. Customary law. Impact of social norms generally ignored by economists. ILMs allow firms to establish a clear set of rules around pay and promotion. This creates both stability and a sense of fairness (so popular with employees).

(Plus, tournaments – the competition for promotion – a very effective way to incentivise effort)

Notes:

2 is Related to 1. Even if the skills needed by the firm are specific, so there is a reduced incentive for individuals to pay for it, in many cases it is not just a matter of who pays for the training (as in the above) but how they are trained. In many cases on the job training will be essential, and so again firms need to construct their own labour markets

In sum in an ILM the central characteristic is that rather than the pricing and allocation of labour being done competitively on the open market, what employers do is to decide the price and allocation of labour according to a set of rules and procedures.

They argue that ILMs are generally popular with employees, but can also add value in certain circumstances to employers.

Societies have two basic mechanism for dealing with social or economic problems:

1. Exit is the neoclassical economists’ mechanism of choice: dissatisfaction communicated through the (silent) loss of customers and revenue. Cut and run.

2. Voice is political mechanism - to complain about the good or service, more informative option and one more likely to lead to repair. Stand and fight.

For example: Citizens can emigrate or protest, employees can quit or voice concerns, consumers can shop elsewhere or ask to speak to manager.

Loyalty: Can influence the cost benefit analysis of whether to use exit or voice. H also argues loyalty increased where voice is allowed.

Notes:

The basic concept is as follows: Members of an organisation, whether a business, a nation or any other group, have two possible responses when they perceive that an organisation is showing a deterioration in the benefits they offer. They can either:

Exit (withdraw from the relationship)

Voice (attempt to repair or improve the relationship)

For example: citizens can emigrate or protest. Employees can quit or express their concerns. Customers can shop elsewhere or ask to speak to the manager.

Exit is associated with the market mechanism (and Adam Smith’s invisible hand)

Voice is associated with the political mechanism and can at times be confrontational.

While both exit and voice can be used to measure decline in an organisation, voice is by nature more informative in that it also provides the reason for the decline.

Exit and voice also interact. Voice, by proving greater opportunity for feedback and criticism can reduce exit. Stifling of dissent leads to increased pressure to use exit.

Loyalty can influence the cost benefit analysis of whether people will use exit or voice. Where there is loyalty, exit may be reduced. Especially where the options for exit are not so appealing. H also argues that loyalty will be increased where members are allowed to exercise their voice.

EXIT:

-Informationally inefficient: may be costly/ difficult to find out automatically why consumers or employees are leaving.

-Customer or employee acquisition and re-acquisition may be more costly than retention (it normally is)

VOICE:

-By encouraging and providing outlets for Voice, Hirschman argues, organisations can improve their performance and reduce exit.

-Voice allows for more effective feedback to remedy defects more quickly

The most famous book in labour economics and industrial relations

Draws on Hirschman’s concept of voice: The “two faces” of unions:

1. Monopoly face (-ve)

   -Able to raise wages above competitive levels

   -(They may hold back to keep employment high)

   -If union operating in a perfectly competitive market and all they did was to raise wages, then very quickly firms with unions would go out of business.

   -Union monopoly power is likely to be closely related to the market power of the sector it organises.

2. Collective voice face (+ve)

   -Trade unions can be seen as a vehicle for collective voice

   -Unions can channel worker discontent to improve conditions (and so – potentially – productivity)

   -They reference Hirschman: Exit, Voice, and Loyalty

   
   

   Notes: Freeman & Medoff note the basic dispute has been between Managers who complain about inflexibility of unionisation (at best) and at worst crime riddled elitist, non-democratic institutions versus those who believe unions are a force for good. The go on to argue that “industrial relations experts have long stressed the ways in which collective bargaining can induce better management and higher productivity”

   
   

   Voice Face can increase productivity:

   -Reduce turnover

   -Investment in skills and training

   -More senior workers -> less rivalry -> more knowledge transfer

   -Pressure on management to be more efficient (tighten standards)

   -Communication channel to increase flow of information

   
   

   
   

   
   

   
   

Chapter 11 presented evidence on whether unions were good or bad for firm productivity.

Chapter 12 on unions and profitability.

F&M concluded that in general unions tend to increase productivity – although this effect varies to no small extent on the context. In contrast unionism almost always lowers firm profitability.

Their (optimistic?) conclusion “On balance, unionisation appears to improve rather than harm the social and economic system.....unions are associated with greater efficiency in most settings, reduce overall earning inequality, and contribute to rather than detract from, economic and political freedom.” (p.19)

Notes:

In chapter 11 They outline three routes through which unions impact productivity:

One is that given union wage gains, firms may shift towards more capital and higher quality labour.

The next two routes, one depresses the other raises productivity:

- Restrictive work rules (including inefficient staffing requirements and limited incentives or restrictions on management discretion) depress productivity

- Voice effects leading to reduced quitting and improved personal policies may lead to increased productivity (but they emphasise that this relies on good labour relations).

Hirsch (2003) notes that empirical evidence on unions and productivity was rather sketchy in 1984, and it remains less clear cut today. H also argues that F&M, despite showing rather mixed evidence on the impact of unions on productivity, claim that their research suggest that unionised establishments are more productive that non-union establishment.

F&M then put forward various observations/ explanations for why or how unions impact productivity. They suggest the following reasons most of which fit into their collective voice argument: lower quit rates, more professional managers.

Ultimately F&M argue that “unionism per se is neither a plus nor a minus to productivity. What matters is how unions and managers interact at the workplace (p.179).

In chapter 12 they summarise what was then very limited evidence on unions and profitability. They conclude that unions lower profitability, and Hirsch (2003) argues that subsequent work fully supports this conclusion.

Monopoly institutions reduce society’s output in three ways

1. Union-won wage increases cause a misallocation of resources (fewer workers, more capital & higher skilled workers)

2. Strikes reduce GDP

3. Work restrictions lower productivity

Voice face can increase productivity

1. Reduce turnover

2. Investment in skills and training

3. More senior works, less rivalry -> More knowledge transfer

-> Also adverse effect that less qualified workers get the jobs

4. Pressure on management to be more efficient (tighten standards)

5.Communication channel can increase flow of information

F&M now contrast both faces – The findings

-Higher wages & Lower inequality among workers

-Reduced turnover

-Higher productivity

-On balance, unionization appears to improve rather than to harm the social and economic system

-Greater efficiency in most settings

-Reduce overall earnings inequality

-Contribute to economic and political freedom + Democracy

-Less profitability

-Union workers are more dicontent

-> The voice face of unions dominates the monopoly face

## **Union Voice Benefits for Employers - My Answer:**

**The argument has merit but comes with important caveats.**

---

### **YES - Unions CAN Benefit Employers Through Voice:**

**Freeman & Medoff's "Two Faces" (1984):**

**Positive "Voice Face":**

- Unions channel worker discontent constructively rather than through exit (quitting)

- **Exit is costly** - losing workers means recruitment, training costs

- **Voice is cheaper** - fix problems, retain workers, improve productivity

- Provides organised feedback mechanism to identify and remedy issues quickly

- Can actually **increase productivity** in unionised establishments (their contested finding)

**Additional employer benefits:**

- Union wage premium forces firms to shift to higher quality labour and more capital (efficiency gains)

- Structured communication channel - one negotiation vs. many individual complaints

- Reduces turnover, builds institutional knowledge

- Can prevent wildcat strikes by channeling grievances through formal processes

---

### **BUT - Significant Downsides for Employers:**

**Negative "Monopoly Face":**

- Wage premium above market rates (cost increase)

- Restrictive work rules can depress productivity

- Reduced management flexibility

- Joint decision-making (loss of unilateral control)

**The employer dilemma:**

- Benefits of voice < costs of wage premium + restrictive practices?

- Employers may prefer to "make" voice themselves (non-union mechanisms) rather than "buy" it from unions

---

### **What the Evidence Shows:**

**Mixed voice works best at firm level:**

- "Mixed" (union + management voice) > "Direct" (management only) > "Union only"

- Suggests unions are valuable BUT work best when complemented by direct communication

**BUT employers increasingly choose non-union voice:**

- UK data shows union-only voice declined most

- Non-union (management-initiated) voice expanded

- **Very few firms try to manage without ANY voice** - they just prefer to control it themselves

---

### **My Assessment:**

**The argument is theoretically sound** - unions DO provide valuable collective voice that can benefit employers by:

- Reducing costly turnover

- Providing structured feedback

- Improving working conditions (which can boost productivity)

**However, in practice:**

- Employers rationally prefer to "make" their own voice mechanisms to avoid wage premiums and loss of control

- This explains union decline even though firms clearly want voice

- The "voice benefit" exists, but employers have found cheaper ways to get it

**Bottom line:** Unions' voice function IS valuable, but not valuable enough (from employers' perspective) to outweigh the costs. That's why we see the shift toward employer-controlled voice mechanisms rather than union voice.

In 1914 LE established an advisory board, it met twice a month and provided a forum in which employees could bring issue of concern to management attention. All advisory board meetings were chaired by the chairman or president to LE, and issues were either resolved on the spot, or assigned to an executive. Minutes of all these meetings were posted on bulletin boards. Workers felt that the advisory board was a way of getting immediate attention for their problems – although it was clear management had the final decision. But if you did not get the decision you wanted, you could raise it, and a clear answer would be given why they had decided against it.

The advisory board has nudged the company into many progressive policies such as group life insurance in 1915 and paid vacations in 1923. According to Frank Koller’s book Spark it also helped to keep LE union free by the voice it gave to workers and the policies it inspired.

Senior management operated an open-door policy, and the CEO interviewed said that workers took the opportunity to talk to him at least twice a week.

OECD report: notes that union voice alone did not seem to be associated with better conditions.

-Olsen’s theories considered the economic costs of collective action (and the distinction between public and private goods)

-Argued people will only engage in collective action where (private) benefits of doing so exceed (private) costs.

-Costs may be financial (union subscriptions) and non-financial (time)

The central problem is one of freeriding

-When the benefits to collective are “public goods” (available to all – whether or not they have paid for them – ‘non excludable’ in economic jargon) then free riding is likely. This is a big problem for unions

Solution 1: Private goods (goods only available to those who have paid for them) e.g. representation in grievance procedures

Solution 2: “Special interests”. Olson suggests unions may also form around highly motivated individuals driven by ideological beliefs with non-financial concerns about fairness (re Amazon case)

Notes:

https://www.economist.com/obituary/1998/03/05/mancur-olson

Olson (1971) – his theories shed light on the economic costs of collective action (and the distinction between public and private goods). Olsen argues people will only engage in collective action where the (private) benefits of doing so exceed the (private) costs.

The costs he considers may be financial (union subscriptions) and non-financial (time).

The central problem that Olson notes is one of freeriding. When the benefits to collective are ‘public goods” (available to all – whether or not they have paid for them – non excludable in economic jargon) then free riding is likely.

The implication is that for collective action to get off the ground, the union must also provide ‘private goods’ (eg goods restricted to only those who have paid for them). He uses the term selective incentive to refer to private goods.

Relating this to Industrial relations. In Olsen’s view unions cannot only be involved in collective bargaining (this is a public good – non-union members get the same pay rise as union members). They must also provide private goods – such as representation in disciplinary and grievance procedures. Where possible unions will try to make unionisation compulsory or encourage employers to advocate or facilitate union membership.

Problems with Olson’s approach: there seems to be no good fit between bouts of union membership and the existence of special conditions.

Unions provide 3 kinds of service:

1. Represent workers collectively in bargaining over wages and conditions (public goods)

2. Represent workers individually in disciplinary procedures (employment contract is incomplete) (private goods)

3. Traditionally they offered personal insurance against misadventure such as unemployment or ill health (welfare state has now largely taken this function) (private good)

   
   

Why do managers choose to deal with unions?

1. Because economic benefit outweighed the costs (Bargaining power argument; Strikes are costly)

2. There are economic rents that managers can share to make their job easier

3. What if unions could be managed to provide benefits to the firm (Switching costs are high)

When is unionization more likely?

1. Perishable products

2. Natural monopoly. Where there is a monopoly, unions have substantial power to inflict damage on the consumer

3. State ownership: Unionisation is much higher in the public sector

4. When skill level is high (harder to substitute employees in a conflict)

5. Occupational communities: some industries have historically been more conflict prone than others

   
   

What caused union decline in the late 20th Century?

The welfare state started to take on some roles that were previously carried out by unions

Employer want voice but not necessarily wage premiums -> They invest in their own forms of voice (Inconsistent with Taylorism)

The state of voice mechanisms

Prominence of

1. Non-union only voice

2. Dual voice

Freeman & Medoff (1984)

Found a union wage effect of 10 – 15% (looking at earlier data)

Using their own data (US) they found an even higher union premium of 20-30%

But there was considerable variation – related to wage sensitivity of the demand of labour

OECD (2019)

United States and the United Kingdom - union membership wage premium of between 10% and 15%

Blanchflower & Bryson (2004)

2004: Union wage gap still exists although it has declined (due to increased competition)

In 2000 was about 17-21% in private sector and about 14% in public sector

Quality of work environment better on average in countries with well organised social partners and union collective agreements

BUT: At firm level, quality of working environment higher in firms where there is “direct” or “mixed” forms of voice, compared to union only voice

“Direct” = No union, i.e. management-initiated voice mechanisms

“Mixed” = Union voice plus management-initiated voice

(In this case – they caution that correlation does not equal causation)

Evidence wage premium and benefits spill over to non-union members

And this last one presents a problem…..

Willman 2014:

More non-union only voice and dual voice

Freeman & Medoff’s central argument was that employers may also reap the benefit of union provided voice

Given unions increase wages, they suggest three effects (two positive, one negative):

1. Due to union wage premium, firms shift to more capital and higher quality labour (positive)

2. Voice effects (potentially positive)

3. Restrictive work rules depress productivity (negative)

Their data, they argued, showed: “productivity is generally higher in unionised establishments than in otherwise comparable establishments that are non-union”

Note this has been challenged – See Hirsch

Middle-class formation: Historical role in economic development (Di Maggio, 2001)

Political stability: Channeling discontent through institutions https://www.epi.org/publication/unions-help-reduce-disparities-and-strengthen-our-democracy/

Innovation pressure: Forcing management to innovate beyond cost-cutting (Hall & Soskice)

Notes:

DiMaggio argues that in many cases the state (which Marx had dismissed as the “executive committee of the ruling class”) passed protective legislation and in many cases supported unions. In other words, European social democracies increased the scope and generosity of welfare, finding a “politically stable trade-off between growth, profits, and welfare anchored in an alliance between the liberal state and the bureaucratic, oligopolistic firm, with trade unions as junior partners.” (p.12)

For Workers:

Union dues – membership fees (economic cost)

Unemployment risk – higher wages mean firms hire fewer workers

Strike costs – lost wages during industrial action

Less individual flexibility – can't negotiate own deals

Seniority systems – frustrates talented younger workers who can't advance quickly

For Employers:

Higher wage costs – union wage premium (15-20% typically)

Lower profits – unions "tax" company returns

Reduced managerial discretion – restrictive work rules limit flexibility

Slower decision-making – everything must be negotiated

Strike threats/disruption – operational uncertainty

Slower growth – lower profitability → less investment in expansion

Administrative burden – time and resources managing union relations

For the Economy/Society:

Deadweight welfare loss – monopoly face distorts resource allocation

Reduced employment – higher wages in unionized sectors = fewer jobs

Insider-outsider problems – unions protect existing members at expense of unemployed

Less labour market flexibility – harder to adjust to economic shocks

Strike costs – lost output for economy

Potential inflationary pressure – wage spirals

Competitive disadvantage – unionized firms vs non-unionized competitors

The Empirical Evidence (from your readings):

Freeman & Medoff originally argued productivity gains offset costs

But Hirsch (2004) challenged this:

Average union productivity effect ≈ zero (not clearly positive)

Union effects on profits clearly negative

Slower growth and investment in unionized firms

Monopoly face has dominated in practice

Key Debate Point:

Do the voice benefits (reduced turnover, better communication, shock effect) outweigh the monopoly costs (higher wages, restrictive rules, reduced flexibility)?

Current consensus: For most firms, costs exceed benefits – which explains union decline in private sector.

A response to Freeman & Medoff – they revisit the argument and data

If a union achieves a wage premium are there and compensating positives?

Commom argument in the union literature is that wage premium provides an incentive for employers to upgrade the skills of their workforce

-> Wessels 1994: Unions could then demand even higher wages; Employer might anticipate and not upgrade

Overall, Hirsch argues the evidence produced since WDUD suggest that F&M’s characterisation of the effects of unionisation on productivity was overly optimistic

Raises the questions if the two faces of unions are truly distinct

-> Sufficiently accurate description for union’s activities while being sufficiently broad to include a wide range of union effects

-Not distinct nor without ambiguity

1. Workers preferences and demands depend in no small part on union bargaining power (M or V?)

2. Management response to unions influences workplace outcomes (M or V?)

3. Workers have a desire for greater voice in the workplace

Productivity: Union impact is close to zero and likely to be somewhat negative

Profitability: Subsequent evidence points unambiguously to lower profitability among union companies (typically 10-20% lower). Results hold for a v ariety of different profit measures / time periods / industries

Investment in capital: typcal unionized firm has 6% lower capital investment than its observationally equivalent non-union counterpart

Business survival: US unions are associated with slower employment growth, but little or no differences in rates of business survival or failure (unions willing to drive a business to the edge of a cliff, but not over it?)

Putting Freeman & Medoff 1984 into context:

-They overemphasized the positive effects of unions on members, economic performance, and society

-A major goal of WDUD was to critique neoclassical economics for their then disproportionate focus on the „single face“ monopoly aspect of labour unions

-They emphasize that there is a second and perhaps primary face of unions – the collective voice/institutional response face

-Their prediction that productivity effects are larger in more competitive environments appears to hold up

Methodological Challenges in Studying Union Effects

Most studies attempt to compare various outcomes such as productivity, profitability between firms or industries with different levels of unionisation, whilst controlling for other measurable factors. BUT

Problem 1: Omitted Variables

For example, if older factories have lower productivity, and also union density is higher in older factories, then any estimates that fail to control for age will mean that part of the effect of age on productivity would be included in a (biased) estimate of the effect of unions on productivity.

In simple terms:

• Older factories → lower productivity (due to old equipment)

• Older factories → more likely to be unionised (historical reasons)

• If you don't control for "age", you wrongly blame unions for the productivity problem that's actually caused by old equipment

Problem 2: Union Status Not Random (Selection Bias)

Unions will likely only survive in firms (or industries) with high potential profits. In this case, estimate of the impact of unions on profitability (which are typically negative) would be an underestimate.

In simple terms:

• Unions only form/survive where firms have profits to share

• These are already the most profitable firms

• So when we see "unionised firms have lower profits", the real effect is even worse than it looks

• Because without unions, those firms would have had even higher profits

Problem 3: Generalization problem / Longitudinal Data (Causality Issues)

For example: changes in productivity before and after changes in unionisation. But change in status unlikely to be random, and changes in productivity take time - whilst there are likely a whole host of other changes in the economy which mean causality hard to prove.

In simple terms:

• You could track a firm before and after unionisation

• But firms don't randomly become unionised - something triggered it (maybe already declining productivity)

• Productivity changes take time to show up

• Meanwhile, the economy is changing in other ways (recession, technology, competition)

• Very hard to prove unions caused the productivity change

The Bottom Line:

Measuring union effects is really difficult because:

1. Omitted variables bias the results

2. Selection effects - unions don't form randomly

3. Causality is nearly impossible to prove definitively

This is why economists still debate whether unions help or hurt productivity, even after decades of research.

Union Busting Consultants

A handful of law and consulting firms to keep companies' union free:

The three biggest law firms that do so are: Littler Mendelson, Ogletree and Jackson Lewis

Consultants include IRI and the Labour Relations Institute https://lrionline.com/

LRI: 'We've helped thousands of companies stay union-free'

'The best campaign is the one that never happens'

Ikea: turned to Ogletree Deakins to help it crush unionisation efforts in Stoughton, Massachusetts store

Google hired IRI consultants for advice on how to deal with growing employee unrest

Avoiding Unions 101

Hard to get good data (many tactics unlawful)

(1) Monitoring unrest in the workplace

(2) Union inoculation

(3) Captivating workers

Walmart the US's largest employer (2.3m employees worldwide)

• Has successfully avoided unionisation

• But faced a "torrent" of lawsuits for its anti-union activities

The best response?

'A company gets the union it deserves'

Letter from CEO of Costco:

"The fact that a majority of Norfolk employees felt that they wanted or needed a union constitutes a failure on our part"