Download as pdf or txt
Download as pdf or txt
You are on page 1of 3

T3.

Firm’s Size

1. Data of Firm’s Sizes


Methodological issue: how do we define ‘size’?

• Physical measures clearly do not make sense


• Output value varies a lot from sector to sector
• The same can be said of balance sheet variables.

Employment levels are the best choice

• Easily observable
• Correlated with any other sensible measure (if there’s growth more people will be hired)

But:

• Self-employment means that numbers of firms with L=1 or L=0 are unreliable.
• Outsourcing leads to smaller firms when measured by employment. We may see that firm
sizes are decreasing not because they are smaller but because they are outsourcing.

LARGE NUMBER OF VERY SMALL FIRMS

94% of firms are micro, only 0.11% are large.

40% of employment in micro firms, 26% in large ones.

Similar firm size distribution in Italy or France, but larger firms in Germany.

2. Explanations Of Differences in Firm Sizes


Technical: Different sectors employ different technologies, leading to different ‘minimum efficient
scale’. If countries specialise in different sectors, different average firm sizes will result.

• Examples: bars and restaurants have an efficient scale smaller than that of a chemical factory.
A country specialized in tourism will have smaller firms.

Therefore, Firms in the same sector should be of similar size across countries.

Institutional: The size of a firm results from decisions related to its ‘vertical integration’ (make or
buy?), which depend on transaction and contractual costs.

• Examples: contract enforcement by the judicial system, credit availability, tax regulations, etc.
Some lead to larger firms, other to smaller ones.

→Firms in the same country should be of similar size across sectors.

Organisational: related to entrepreneurial & management quality. Maybe the managers are not good
enough in order to run a big company but this is very difficult to measure.

COMPARISON OF FIRM SIZES ACROSS SECTOR AND COUNTRIES

The table shows the average firm size per sector and country. They took 15 EU countries and made
this list of market-based sectors (private sector).

• EU15(first column): average employment. Because of the numbers we can assume that the
very small firms are being dropped out.
• Capital intensive activities have larger firms.
• Country cols: relative size with respect to EU15=1.
• Large intra-sectoral differences (rows): if the technical explanation was the only reason, these
should be all ones.
How much of the difference in firm sizes is due to sector specialization (technical explanation) and
how much to national (institutional explanation) effects?

𝑠𝑗 − 𝑠̅ = ∑ 𝑤𝑖𝑗 𝑠𝑖𝑗 − ∑ ̅̅̅


𝑤𝑖 𝑠̅𝑖 = ∆𝑤 + ∆𝑠 + ∆𝑤𝑠
𝑖 𝑖

= ∑(𝑤𝑖𝑗 − ̅̅̅)𝑠
𝑤𝑖 ̅𝑖 + ∑(𝑠𝑖𝑗 − 𝑠̅)𝑤
𝑖 ̅̅̅𝑖 + ∑(𝑠𝑖𝑗 − 𝑠
̅)(𝑤
𝑖 𝑖𝑗 − ̅̅̅)
𝑤𝑖
𝑖 𝑖 𝑖

𝑠𝑖𝑗 : average size of firms in sector i in country j

𝑤𝑖𝑗 : share of employment in sector i in country j

𝑠̅,
𝑖 ̅̅̅:
𝑤𝑖 sector level EU averages

∆𝑤 : difference due to sectoral composition (technical explanation)

∆𝑠 : difference due to size differences within sectors (institutional explanation)

∆𝑤𝑠 : interaction term

Spanish firms are on average 42% smaller than EU ones.

• 6% of that difference is due to specialisation in sectors with smaller firms


• 39% due to within sector differences (smaller firms in the same sectors)
• -3% due to an unexplained interaction term

In all countries (but Sweden) institutional reasons are the main sources of observed differences in firm
size.

3. Firm Size and Productivity


There are arguments to explain causality in each direction:

- Size increases productivity (learning by growing)


o A large firm can benefit from economies of scale (AC decreases when you increase
your level of production) and scope (AC decreases when you diversify your
production), be able to invest in new technologies, coordinate more resources, learn
about more efficient methods, etc. When firms become bigger they are going to be
more productive.
o If this is so, we need to design policies addressed at increasing firm size, and
productivity increases will follow.
- Productivity increases firm size (growing by learning)
o When small firms increase their productivity, they grow. Or, in an opposite sense,
firms that don’t increase their productivity disappear and don’t ever grow. This is
related to Schumpeter’s idea of ‘creative destruction’ and small firms’ higher
incentives to increase their productivity. Industrial dynamic is by having a small firm
enter into a given market, the ones that succeed become market leaders. Increasing
productivity depends on small firms making the effort to grow.
o If this is the right explanation, then growth-support policies will be useless, and may
have negative effects. Productivity increases have to be encouraged in other ways.

Moral-Benito (2018) tested empirically the validity of each of the two hypotheses for Spain.

Individual-level data from the Balance-Sheet database (Central de Balances) of the Bank of Spain.
Compares well with the Business Register (Registro Mercantil – the source that feeds SABI & BVD).

TFP per firm estimated from 𝑦𝑖𝑡 = 𝛼𝐿 𝑙𝑖𝑡 + 𝛼𝐾 𝑘𝑖𝑡 + 𝛼𝑀 𝑚𝑖𝑡 + 𝛼𝑖𝑡

TFP difference between a large and a micro firm is 623%: 𝑒 (2.71−0.88) = 𝑒 1.83 = 6.23. This means that
on average, bigger firms are 6.23 times more productive in terms of TFP.

The differences hold across sectors.


The distribution of productivity levels for all firms shows wide differences between firms of the same
size. This justifies the need to use individual -level data in empirical analyses in order to understand
what’s causing productivity at each level. There is a difference in productivity and at data level.

Four estimated equations:

a. TFP shock → L growth. What happens with employment when there is a shock=
b. TFP shock → cum. L growth.
c. L shock → TFP growth: What happens to TFP when there’s a labour cost
d. L shock → cum. TFP growth

Productivity shocks lead to employment growth during the 3 next years, with lasting effects 5 years
later. Size shocks increase productivity only on the following year, with no lasting effects.

These results provide support to the ‘growing by learning’ hypothesis. However, this doesn’t mean
that it’s exact for all particular firms.

You might also like